The Under Net & Being Data Led with PMG's Jason Hartley

When Steve Jobs held up the first iPhone, no one thought New York taxis would be in trouble. But the speed and complexity of technological development has now got us demanding Ubers when and where we want them - a lifestyle change that points to how early we may be in the way AI is developing and influencing our world, our search patterns, and our consumer behaviour.

Jason Hartley, Head of Media Innovation and Trust at PMG, discusses the importance of data-informed marketing over data-led approaches, emphasising the need to challenge assumptions and consider human behaviour. "Consumers are moving faster than we can measure them," but we have correct instincts about that. So, instead of relying only on data, use those gut instincts and then look at the data points to confirm and test.

As his shiny new object, Jason introduces the concept of the "Under Net," describing the invisible infrastructure and AI layers shaping online experiences. He also addresses the commercial and environmental implications of AI, suggesting that while monetisation challenges persist, the technology's potential for efficiency and utility is significant.

Tune in for lots of interesting data driven (or anti data driven?) marketing tips and a reminder that "more information, more data doesn't necessarily mean a happier, better world."

The following gives you a good idea of what was said, but it’s not 100% accurate.

Jason Hartley 0:00
I hear the expression data led a lot, and I say it, and people say it, and I think that's kind of a mistake. I think it needs to be data, informed, data, that challenges the assumptions that are most closely held. They're usually ignored, but that's probably when they're most valuable.

Speaker 0:21
Are you a fitness loving foodie? Then defy snacks is for you. Defy offers real high quality chocolate you love, combined with functional ingredients, giving you as much as 26 grammes of protein per bag. So you can be both a little bit bad and a little bit good at the same time as a female owned brand, defy dedicates a portion of its profits to other women owned businesses currently sold in the US. You can find us on defy with an i snacks.com and follow us on social at defy snacks.

Tom Ollerton 0:53
Hello and welcome to the shiny new object podcast. My name is Tom Ollerton. I'm the founder of automated creative, the creative effectiveness ad tech platform, and this is a podcast about the future of data driven marketing. Every week or so I have the pleasure and the privilege of interviewing one of our industry's leaders about their vision for where this industry is going. And this week is no different. I am on a call with the lovely Jason Hartley, who is head of media innovation and trust at PMG, So Jason, for anyone who doesn't know who you are and what you do, could you give us a bit of background?

Jason Hartley 1:27
Sure. Well, first of all, thanks for having me. It's very fun to talk about these topics. So this is great, but I should start with my title, because it sounds kind of made up and grandiose. But so media innovation and trust. Basically it is we're doing, trying to do cool, new things, but doing them the right way. You know, we're thinking about privacy, we're thinking about the ethical implications of AI and so on. And I feel really lucky to have that position, because, you know, for a lot of my career, it's been, can we do this, not should we do this? And it's nice to see the industry moving along. As well as the company I work for, I've been in paid search for about 20 years. I've been focusing on privacy the last several years. In my spare time, which is vanishingly small, I make music and I do yoga and lots of other things. Sort of try to take all of those experiences in my life and put it together and think about what the future might look like for for us marketers, and maybe beyond.

Tom Ollerton 2:27
So what has been that moment in your career where you were incredibly red faced and you wanted the world to swallow you up, but in retrospect, you're quite glad that it happened?

Jason Hartley 2:27
There's so many of those that I could point to, and of course, that's usually that's when you grow, when you learn, is when you make the biggest mistakes. But the one that always comes to mind to me was actually something that felt like a huge success and then turned into a failure. Years and years ago, at another job, I got the sense that we really understood the way paid search worked. We understood the rules. We understood what Google and Microsoft were telling us was what was the right thing to do. And it felt like, well, if I could describe this, then we could code something to see, were we doing it right? I was able to get a team together and pitch this to our CEO, and he was all into it, and we made this incredible product that would predict performance on your paid search account. This is long before opt to score and things like that. So there really wasn't a product out there that was widely known. I'm sure other people were working on it, but at any rate, I've took very great care to make sure that it looked good. I worked with our product developers and just said, like this should be we have to think about this as being an environment that people are going to spend hours in. So it should feel wonderful. They should get some pleasure out of it. It should work nicely. It should look beautiful. The thing that I thought about was, OXO Good Grips the way it feels to use their can opener compared to what it usually felt like to use a can opener, but any rate. So we did all that, we got really good feedback about it. And like I say, it felt like a real success, because it actually, when we tested, it worked out really well if it predicted that something was going to improve performance. And we did it, it actually did, but we missed something that ended up being fundamental, and that was it wasn't connected to the to AdWords at the time, so you couldn't actually make changes in this tool. You had to have your two monitors open and move from one to the other with your eyes. And it turned out that people don't want to do that even though it works. They didn't want to do it because it wasn't seamlessly integrated. And I pushed and I pushed and I prodded and I cajoled and I did all the things that I could as a department head to try to get people to adopt it. And very good people, very smart people, just wouldn't. Never use it. And so all of this time and development and all these things, and we put it in pitches and everything, and it ended up just being an absolute flop, even though it worked. And so it's useful, because I learned a lot of things there. One was, you can't make assumptions, that's for sure. We did a lot of testing with people, and they were, they were on board, but we didn't ask always the right questions. We probably led people down the path. They probably didn't want to tell me, No, I don't like it, or it lacks this, or so on. So a little bit of humility and self awareness, I learned from that, but also just human nature. We have to think about the fact that, you know, people don't want their lives to be more complicated. They also don't want to feel like they're that someone's watching them. You know that this was a dashboard that I could look at as a head of department and see what accounts were doing what and even though I wasn't doing that to catch anybody, it was just so that I could easily tell our CEO this is how we're doing as a department. So at any rate, it was a total flop, but it was really good for me to think about, okay, well, when I think about things that are happening right now and try to extrapolate out into the future, you know what's going to work? What's going to be something that's going to connect with an average consumer who doesn't need a more complicated life, and they don't care about AI and they don't care about agents or any of these things, they just care about getting the thing that they want, or having a good experience, or seeing something that gives them pleasure, or whatever it is, and so all of that sort of wrapped into a worldview of human nature is undefeated. You know, I can tell you it's amazing, I can show you it's amazing. But if I'm not appealing to the humanity of someone and really thinking about their lives, then I'm probably going to miss some connections that I could have made otherwise.

Tom Ollerton 6:48
Yeah, my thing for that is being right is not enough. I say it's you can be as right as you like, but um, was it? Was it? What? What did it? Was it? Thomas Aquinas say that convincing untruth is better than a unconvincing truth.

Jason Hartley 7:05
It was crazy. I mean, I could show that it actually worked. It was really strange, because we are data driven people, you know. And it just seemed to me like, Well, gosh, this was a no brainer. Of course you would use this. But when it's Tuesday afternoon and you have to open up yet another thing, and you're trying to do things quickly, and you have to drill down into these things and find where the problems are, you know, you just people just not going to use it. And doesn't mean they're not smart, doesn't mean they're not dedicated or anything like that. It just means you've made a product that doesn't solve their needs in a way that they needed to be solved.

Tom Ollerton 7:38
Yeah, and the problem you actually sold is your desire to be innovative, right? And sometimes that means you also have solved someone's problem. And boy, have I done many things in the count of like, well, I thought it was cool, so that's a lovely story. So have you got a specific bit of advice that you would give to someone to become a better data driven marketer?

Jason Hartley 8:02
Yeah, I think so. Again, I bring so much humility after that experience and so many other ones that are probably more embarrassing, that one probably made me sound smarter than I really am. I could probably do some ones that really are embarrassing, but at any rate, I would say, look beyond the data, to be a better data driven marker. Or, you know, I hear the expression data led a lot, and I say it, and people say and I think that's kind of a mistake. I think it needs to be data informed instead of lead, because there's a larger world outside the data. And I think that sometimes we can measure ourselves into a corner, and get into that old problem of just thinking that if you can't measure it, it doesn't matter. And I think that as we see user behaviour and consumer behaviour changing so much and so rapidly with everything that's going on with AI, they're moving faster than we can measure them, and so we can decide, well, we're going to ignore it, because we don't have the data similar to what we did with mobile for years. Or we can say these, this is where consumers are, this is what they want. We need to be there. Now, let's try to think about, you know, signal, not certainty. What can we learn from what we have and make good decisions that way? But I think there is a, there is a, you know, there was almost a cult of data years ago where it felt like everything was an engineering problem, but we're dealing with people, and we're trying to connect with them in meaningful ways. And the data can help, for sure, but it shouldn't lead you. You know, I think that there's a combination of things that you have to put together, and then ultimately you put together what you're good at as a human with it, which is just like we have instincts that are right, but we need to look at data to confirm those. We also have to look at our biases. We have to look at the biases in the data and so on. So I think it is you don't just trust the data, but you do use it as a tool. And, you know, look at the data that challenges your assumption probably the hardest, because what, in my experience, it does seem that data that challenges the the assumptions that are most closely held, they're usually ignored, but that's probably when they're most valuable.

Tom Ollerton 10:15
So going to move on now to your shiny new object, which is the Under net. And we've had 300 plus episodes of this podcast, and very rarely did people choose exactly the same, but we've definitely not had an under net before. So what is that? What is the Under net? And why is it your shiny new object?

Jason Hartley 10:33
So the under net was something that we were talking about internally as we were thinking about the fact that increasingly the online experience is being powered that that by things that are kind of invisible to people. And so as we started to think about the emergence of agents and how llms work to sort of organise and summarise the world, and then give you what looks like a baked answer. It felt right, because the internet is there, and we live on it, and it's a surface level thing, and we don't really think a lot about what's going underneath. And so a lot of the pipes and things that we had built underneath there, the infrastructure of the internet, is now being used in a new way that's not really visible to us, we have these new, invisible influencers in our lives that we don't really recognise. And that's really the under net. It's all of those things together. And what's been interesting in just in the last few weeks, with AI browsers kind of popping up, that's almost I might have to have another name for that. We're considering the overnet for that, because there's this AI layer on top of what you do. So it's like this overlay. So you have this under net, underneath the Internet where all these things are happening, and machines are talking to machines, and they're influencing your experiences. You have the Internet where we're we're doing what we always did. And then there's the over net that has this assistant quality, where you can is, you know, you're not bound by the fact that you're in a website. You can go anywhere in the world while still being on that website. So it's a pretty profound shift, but the undernet is something we've coined, so we'll take a little credit for that. But I think, you know, just speaks to this larger movement where, you know, agents are more a part of our lives and so on. I know everyone talks about agents and what they can do, and there's a lot to be learned there. I think that we're a ways away. But as it is right now, the original idea about the undernet was just the fact that, you know, if you go to chatgpt and it doesn't have the answer in its data, it searches for you, and so it searches instead of you, and that's that's a profound shift that it's happening because you don't really see all the steps that it's taking, or you don't care about them so much, and you're just given a completed answer.

Tom Ollerton 12:56
This episode of the shiny new object podcast is brought to you in partnership with Madfest, whether it's live in London or streamed online to the global marketing community, you can always expect a distinctive and daring blend of fast paced content, startup innovation pitches and unconventional entertainment from Madfest events. You'll find me causing trouble on stage, recording live versions of this podcast and sharing a beer with the nicest and most influential people in marketing. Check it out at www.madfestlondon.com.

Tom Ollerton 13:33
What's your view on the commercials of the all of this? So there's a writer that I... his name completely escapes me. He does a brilliant podcast called Better offline, and he's been railing against Gen AI, what's his name? Ah, it will come back to me. Ed Zitron and great writer, don't know if you know him, and just week after week is his 13,000 words email newsletter talks about how no one's making any money out of this. People have revenue, but I don't know. I butcher these numbers, but Amazon's capital expend, expenditure on AI versus Gen AI, sorry, versus profit. It's like nine to one or something. It's crazy, like, and, you know, Meta up there, and so no one's making any money. So what's going to happen? Because there's yes agents, I get it. Yes, I am a huge user of these services, but, but at some point, if someone's going to get their money back, then they're going to be ineffectively expensive.

Jason Hartley 14:36
Sure. Well, I think that, you know, we have to look at the bigger picture for sure. I think in general, one of the things we're guilty of, and again, kind of going back to that data driven marketer, you know, data is really good at telling you what happened in the past, but it can, it can sort of stick you in the past as well. I mean, we're, we are trying to impose these brand new things onto an existing or onto an existing. Sort of framework that we're comfortable with, I think that there's going to be new ways to monetize this stuff that are going to pop up. There's just too much opportunity here, I think, to ignore. And if you think about, you know, when Steve Jobs held up the iPhone for the first time, there weren't a whole lot of people thinking to themselves, Boy, I bet taxis are going to be in big trouble. You know, they weren't thinking that there was going to be this app ecosystem that was going to make it so that in New York you could get a car instead of, you know, come at when you needed them, rather than calling a car service and waiting for them to have a car available, or standing there in Manhattan at four o'clock, holding your hand up for hours waiting for a taxi to be free. So point is, is, like we're so early here. I think that there's a lot of different ways that this could go. You know, everyone's sort of waiting for, you know, will Open AI have ads and so on. I think that they have bigger aspirations than that. We'll see. I think their subscription service is definitely, you know, that's going to be a steady revenue driver. And to your point, it's maybe not making them any money. But, you know, I look back to what, what Amazon did for years and years they were, they could be a company for a really long time without making a profit, because everybody saw the vision. Everyone believed in it, and everyone could see where it was going. We're not advanced yet enough for that, I think, but everything's much faster now. It feels like that. I think we're... there's a general belief that, you know, Open AI is not going anywhere, and they have some runway to figure this out, to make sure that it's profitable when you so, so at any rate, so I don't see them sort of adopting that model anytime soon, or having to. I think, you know, the smaller players are probably going to have a little bit more incentive to try to monetize it more overtly. But subscriptions definitely are something that people are doing, but it's definitely, I mean, it will get cheaper to process. Models are going to be cheaper to build and so on. So I think, you know, if we, if we look back to the past and we sort of think about the future and how it might mirror that. I think it's going to be a combination of factors where someone's going to figure this out, because ultimately, how are we going to go backward from this? You know, what's going to happen, like is everyone just going to fold the generative AI tent and say, we can't make any money on it, so we're just going to go back. I don't see that happening, and it may be just that it's being built by companies that have really diverse revenue streams and then allowing them to figure out, some time, to figure out how they can actually make money on it. And ultimately, if Open AI decides they can't do it without ads, well, they will have billions of prompts every day. They have hundreds of millions of users already, and they can have extremely relevant ads within their platform for any kind of content that will be very personal and will be intent, will be at the heart of it. Same goes with Google. Same goes with perplexity and grok and all the rest. So if they want to pull that lever, it's there to be pulled. And I think if you do it the right way, that, certainly, you know, it's not, it doesn't take that much of imagination to see how it might look.

Tom Ollerton 18:05
Yeah, there's a few ifs like, you know, the the Amazon model was this all kind of based on the fact that people, people are going to carry on buying more and more stuff online, right? That that was the, you know, that was the easy thing to follow, right? Obviously, in right respect, it was a it was the right horse to back. But I don't really see that with Gen AI in the same way it's like, oh yeah. People use Gen AI to sort of write emails quicker and, you know, like it's and, yes, I have many agents. Well I hate that expression, but I guess I have many gpts or projects, whatever you want to call them. Let's call them agents working. They don't work for me. I make it. It's a computer programme. Let's just call it what it is. We don't need to oversell it. But there's, I wonder, does it really what is the the man or woman on the street thing that they can't possibly live without? I think that's what I've not seen yet. And with your mobile phone, your so your iPhone, example, like the the economics always works, right? Like, we can have an app store and, like, Oh yeah, and so, and yes, you could throw up an app for 3040, 50 grand, or whatever it was, and then you can make revenue from it, whereas, if you're losing, you know, $9 for every dollar you spend. Then, then maybe it's a bit more of a Concord situation. You know, you could fly to New York for three hours from London, but it just didn't stack up financially. So they stopped doing it.

Jason Hartley 19:24
Well, I guess the question is, what do you charge when all the best apps are in your platform? You know, if they, if they actually do, provide a lot of utility, I think what I immediately raised to when I started to think about agents, when they started to emerge, was no one will ever have to overpay again, you know, and that that is something that's pretty interesting, you know, where you can say, I know I want this toothpaste within this certain amount of time, go find it for me, and I never, and I don't have to think about it or expend any effort at all. If you use the AI browser now, you have a window right next to you. So if you go and you find and it navigates. To you, for you on the site, and gives you a bunch of options based on, you know, you're saying, I want the most trendy blah, blah, blah. And then you can say, well, is this the best price? And ask a lot of questions about and I think there's a lot of utility there. But, you know, to your point, I mean, the browser business is not a great business. You know, search has been propping that up in a lot of ways. It's something to watch. But again, I think that we're still so early. I think it's, it's certainly too early to dismiss anything. I think there are little, little bits and pieces to think, okay, you know, it may be more like Napster. You know, where Napster was like, okay, it broke the industry, and they had to figure out how they could do iTunes for 99 cents, and then streaming and so on. And it that was a shift in the business that ended up being okay for the labels, maybe, and for the platforms, but it did hurt artists, you know, because if you're not huge, then it's not great for you. So I think, you know, it could go in a lot of different directions, but I think that there are enough reasons to believe that this is going to stay and provide real value to people, that people will pay for it. And then you throw in the fact that this stuff is powering business as well. And if you look at, you know, all sorts of businesses are incorporating this technology, and they're getting it from the same platforms that we're talking about in a marketing context. That's why I say that the diversity of their revenue could keep them from having to do the types of things that Google and Facebook did when they were needed to make some money and really rely on advertising. They're just more options now that they can have and they can rely on so they might just focus on user experience and making as useful as possible, and then creating an ecosystem where they have developers that share revenue, or whatever it may be. So let's see how that

Tom Ollerton 21:51
evolves, and what's your take on the environmental impact. I'm sure that most, once I hear they'll work out how to make it greener. But there's two quotes that made me think. The first was, every time you use chat GPT, a tree dies, which I thought, and then the other one was to write an email is the equivalent of boiling a kettle? Yeah, yeah, which is a really visceral amount of energy. I don't know how much energy that is, but it's like, if you know, you fill a full kettle and boil it. You know, there's energy's being used there. And, you know, I'm like, Oh, God help me. I don't want to read this brief just, you know, summarise it for me.

Jason Hartley 22:29
Sure, sure. Yeah. I mean, that's definitely, I mean, it's something that we have to figure out. It's something that I've done a lot of presentations about, actually talking about the ethics of this, and that's a big part of it, for sure. I think that I do try to look at the big picture, kind of this is the theme, you know, let's have some humility. Let's not look at just this one data point. What is the, what is the greater context here? I agree that people need to not they need to be aware so they can make better decisions. There are ways to to, you know, batch your your requests and things like that, but if it's a working tool and you're expect to use it, you know, it's kind of hard to be strategic about when you use it, but it isn't that hard to be strategic about knowing that I can just, you know, search this on Google. But of course, if Google is going to take you into AI mode and use the same kind of technology, then you don't really have a whole lot of control over it. But I do think, you know, when I look at the at the full picture, you know, you'll see some people say, Okay, well, you know, you know, to your point, every time someone uses chat, GPT, an angel loses its wings, or whatever it may be there, they'll point out that there are five other things that may have gotten better, or the fact that they can use the same technology to make, you know, transportation more efficient, or factories more efficient, or whatever it may be. So I do think we have to look at the big picture, but at the same time, I think that we all need to keep talking about this and showing how important it is to us. And I think that we can all sort of take our own steps individually, if you're a part of a larger company, like in my role, you know, trust is in my title. So I get to talk to 1000 people about, look, we're doing everything we can as a company, but it will take individuals doing actions as well. But ultimately, it's really going to be up to the bigger companies who are making the biggest dent in this. You know, I think often we talk about environmental problems, we talk about our own personal responsibilities, and sort of leave out the fact that maybe bigger, bigger forces are at work here. And it does have to be a collective action, complete with consumers saying or people saying, you know, this is important to me. You need to fix this and and see where it goes. So it's another thing where it's like, gosh, I would love to come back and talk to you two years from now and see, see where we are there is, is it going to be much more efficient? Are we going to be able to. Have models I saw recently that Deepsea had, I think it was 90% or 93% of Grok's intelligence at 15% of the cost. Is that the future, are we going to be able to do things much more efficiently, much more cheaply and so on? I don't know, but we'll see. But in the meantime, we should all be talking about it, making sure in our own actions. I think we have to be aware of it at the very least, and then make decisions based on what you know.

Tom Ollerton 25:34
So to finish, which under net possibility, capability or agent do you think is being most overlooked by the marketing community?

Jason Hartley 25:46
Interesting, I think, you know, for me, what I keep thinking about is everyone's now really talking about this, and this is something we've had around for a long time. So I you know, and plenty of people are so or have been for quite some time, but when I look at it, I don't hear a lot of conversation about the effects around this. You know about the fact that what we used to do is people are going to be changed, will be a machine doing it for us. And what is the effect does that have on programmatic on the open web. You know, all of these, all this inventory that used to have ads on it, if we're not looking at anymore, what happens to that and so, and also, just like, what does it have? What does it mean for us as human beings? You know, our brains are getting rewired again, and to abdicate the responsibility of doing these things or delegate it is really profound, I think. But we're still, you know, going back to the human nature thing, we still want that joy of discovery. We still want to feel like we've found things, you know, due to our cleverness or our connections or whatever, and and so when I think about this, I can't help but to start panning out and just looking at the entire picture and how it affects it, both in terms of human beings, the way we think and the way we work, the way we feel, the tools that we prefer, the experiences that we like, and how those are going to change. And I think it's really those sort of second order effects that are being not talked about by a lot, because I think people are still in this phase of, well, these, you know, this agentic search, or the AI browser, or whatever, just the actual facts of the thing and the capabilities of it. And, you know, I try to think a little bit beyond that. But if I had to choose one thing that's that's changed, that that, you know, is the thing that I don't hear people talking about it like I thought that they would, which is price comparison. That's, that's the thing that, to me, is just like that was the most obvious use, and I know it's a part of a lot of what we're doing, but the fact that we're empowering people to be able to do that, and not only online, I can use an assistant that's multimodal, and I can go into a flea market and show it this and say, what is that worth? So it's kind of profound to bring that information there, but as we've seen in the past, more information, more data, doesn't necessarily mean a happier, better world. So we'll see how that all pans out. But that was a bit of a long answer, but it's best I got for today.

Tom Ollerton 28:27
Well, Jason, you're going to put a date in the diary for this time in two years, if not sooner. That was brilliant. Thank you. I remember we supposed to have a 15 minute prep call with guests, and we talked about an hour, I think, sort of something on a variety of different topics. But thank you. If someone wants to get in touch with you about the internet or anything else we've talked about, where's a good place and what makes a message that you will actually respond to.

Jason Hartley 28:52
So LinkedIn is the best place to get me. That's something that will definitely get there. And look, if you're reaching out to me because you're genuinely interested, and you you thought that had something interesting to say, or you have something to add, or whatever it is, I'm listening, for sure. I always tell people, you know, I'm not spammable. I'm a voracious reader. I just want to learn, get as many perspectives as possible. If your goal is to try to sell in a service, then I'm probably gonna be a little less excited to respond, unless it is a service that is clearly going to be useful and can maybe, you know, deal with some of the things that we're talking about. So but generally speaking, I'm a nice guy, so you know, if you send me a note, I'm going to respond. But if you have some authentic connection to this, this topic, then you'll be my friend forever.

Tom Ollerton 29:44
Jason, thank you so much for your time.

Jason Hartley 29:46
Okay, thank you. I appreciate it.

Subscribe to our YouTube channel

Listen to our ‘Shiny New Object’ Podcast on Apple Podcasts, Spotify and Soundcloud.