Redefining Marketing in the Age of Web3 and AI: Profound Insights with Ari Lightman, Professor Digital Media & Marketing from Carnegie Mellon University | S3 E22

Ari: 0:00

But in the end, the consumer doesn't care about all that. The consumer wants an organization to understand them holistically.

Joeri: 0:10

Hello everyone. My name is Joeri Billast and this is the Web3 CMO Stories podcast. Our guest today is Ari Lightman. Ari, I hope I pronounced your name in the right way. There's always a challenge. You maybe will have the same with my name. People in the US call me Joeri. So, guys, if you don't know Ari, Ari Lightman is a digital media and marketing professor at Carnegie Mellon University's Heinz College. He's an expert in digital transformation. He leads the unique measuring social course, collaborating with giants like Google and the NFL. Ari has initiated key CMO programs and delves into Web3 analytics research, with a rich background in management, consultancy, and start-ups. He holds degrees from top institutions, including Carnegie Mellon and the University of Toronto.

Joeri: 1:05

This is already a lot, Ari, and, to be honest, there were so many things I could explain to people about you tell people, so I shortened it a bit. If there's something that you want to add, please go ahead. And, of course, yeah, I would also love to hear about you. I have a couple of questions, and the first one would be about digital evolution. How do you see the convergence of marketing and IT evolving now, in the age of Web3? And yeah, because, with all the experience, I'm curious to hear what you think.

Ari: 1:37

Sure, Happy to elaborate on that. In terms of looking at the history of marketing, there was a period of time when it was very much task-oriented. It was a mechanism associated with understanding physical touchpoints with consumers. The data was fairly anecdotal, if you will, and there were mechanisms there where it was more focused on this mechanism associated with how to advertise correctly. A lot of things were very much product-focused looking at specifications, looking at features and then the shift that occurred around market orientation really looked at this sort of evolution associated with let's focus less on technical features and product specifications and more on value definition what value do they provide to consumers? In doing that, especially with the explosion of digital touchpoints that could be everything from customer service departments interacting with consumers on a social platform or a community portal to mechanisms associated with a variety of different types of commerce. Just put a vowel or an acronym in front of commerce as M-commerce, digital commerce, and social commerce. Now there's thing-based commerce, so we have a ton of different types of mechanisms of interacting.

Ari: 3:07

The explosion of data sort of mandated the fact that IT and marketing needed to interact much more proactively and much more functionally, because, with this massive amount of data, how do you store it, how do you collect it, how do you normalize it? How do you create analytical mechanisms on top of it? So this interweaving not just IT and marketing, but also risk and compliance, Because more and more of the data was very much tied to individuals PII information, personal finance information, a variety of different things out there we're collecting to understand and assess what are the needs of different consumer groups, how do we segment them into unique audiences to engage with them more proactively? But what do they need Right, not just now, but in the future? So now you're doing things like predicting and forecasting, trend mapping, analysis, and now, with the explosion of AI, which I'm sure we'll get into, we just have more and more data than ever before. It's multimodal data, so it's coming from multiple different sources interacting with multiple different departments, but in the end, the consumer doesn't care about all that. The consumer wants an organization to understand them holistically, wants an organization to understand them in a mechanism that they feel comfortable with right associated with their information and data, and wants the organization or the brand to buy them value, differentiated value across a multitude of different axes, if you will. So it's becoming more difficult than ever as we're looking at mechanisms associated with automating it around speed, efficiency, portability, and scalability.

Joeri: 5:03

Right, yeah, you mentioned the word AI. AI is everywhere, from analyzing data to creating content. AI and digital media. How do you see AI playing its role, particularly concerning personalized content delivery and user engagement? I'm curious.

Ari: 5:25

Yeah, I mean that's fascinating. Actually, we're talking to a variety of brands right now About how to utilize various different AI technologies to create greater levels of personalization and customization. So for a long period of time, there was a mechanism associated with, to some extent, standardization.

Ari: 5:46

We could scale, we can expand really easily, we could automate, but we live in a highly customized world. My needs are much different than yours, right? So we have to understand that for a brand out there that might have hundreds of thousands, if not millions, if not tens of millions of consumers, their job is to really understand all the different audience segments that are out there, what their needs are, and look at them across a multitude of different types of attributes and associations. So, looking at demographics, of course, which is just foundational, but also psychographics, technographics, and now we're even looking at cryptographics in terms of wallet holding and wallet patterns and all those sorts of really interesting things.

Ari: 6:37

So the question then becomes you know, are you a million listeners out there? Right? And I think after this podcast, you will, I think you will get up. You might think of yourself well, obviously, they're not all the same listener, right? But the question then becomes if there are two unique audience segments, half a million each, or are there a thousand audience segments, right, with a thousand different audiences and listeners each? If you get to a thousand now, you're looking at, I mean, hundreds of associations that you might ascribe to those different audiences, to that different listener segment to understand, assess, and how do you develop, not even customized content to reach them in a way that they can process and they can understand? So you know, one of the things that we haven't mentioned I'm sure we will is, you know, attention span.

Ari: 7:37

You have a very short time period to get my attention. I'm sure there are several audience listeners out there that know you, that know your brand, that listen to you on a weekly or monthly basis and are really fascinated with it right, so they might have sort of you know loyalty.

Ari: 8:00

We're like yeah, I know you're on it. I'm interested in listening to the first. You know a little bit. No, make a decision as to whether I'm going to listen to the whole show or not, and there's a bunch that, if you don't capture their attention, immediately forget it. So the question then becomes as a brand how do we engage with them, how do we reach out to them? How do we establish a relationship with them? And that's really challenging to do when there's a variety of different consumer groups that are out there.

Ari: 8:30

So that's where the idea of customization comes in, and utilizing AI as a mechanism to identify what are those pieces of content. In which mechanism do I deliver it? How do I order it so I get the maximum amount of uptake, right? And for a long period of time, marketers would experiment with this continuously. I remember doing tons of AB testing right, what's the right picture to put in the right post to get the maximum uptake? And then you have everything is sort of control, and then you have one that's variable and then you just do these on a massive level. Ai's ability to do this automatically right.

Ari: 9:17

Identify the right audience. But now you're sort of left with this idea of interacting with the AI system, doing some level of checking and assessment to make sure that it is correct and then it hits the mark with the right efficiency that you're looking at Right. So, even though there are a ton of AI systems out there, from a marketing perspective I would still say that a lot are human assist market assist.

Ari: 9:51

Right, I don't think anybody or any brand out there. Well, there might be some brands that fully automate marketing from an AI perspective. Right, Because in a world where your brand can be hijacked, you can be misinterpreted, and there can be misinformation associated with it. You want to make sure that there's hands on the wheel. Right, yeah, just like a self-driving vehicle.

Joeri: 10:16

Yeah, exactly, you know. I just came back from Dubai, you know, from a Web3 conference, the WOW Summit, and also people were talking about AI and that was actually the message. You know. Someone was even talking about Web 3.5, like Web 3, you know, with everything around NFTs and metaverse and AI, but in Web 3.5, like you would have AI and metaverse coming together like I don't know if you heard already about the term, but I love to hear your thoughts also you know AI in everything that's Web 3 and Metaverse related when it comes to personalizing and so on.

Ari: 10:58

Yeah, I think it's kind of a funny subject in the sense that you know we go through waves, and I think we talked about this before.

Ari: 11:04

Everybody was crazy about blockchain and the pandemic hit. We all lost our collective minds and we're all working remotely. So everybody's like, you know, investing the metaverse thing. This is the next best thing. But then COVID debates somewhat. We all go back to work like, oh, forget about metaverse, let's go to AI. But there's a definite convergence associated with it. If you think about Web 3 and blockchain, one of the really interesting things about utilizing blockchain when it comes to the tokenization of goods and services and utility is now you have a bunch of data that's readily available and accessible.

Ari: 11:46

So where a lot of the brands that I work with we analyze a lot of social data, that's being more of a walled garden is going around it because, you know let's look at Reddit, right Large language models grabbed a bunch of data to train their models on, and now they're saying wait a minute, we need to monetize this information. You can't just grab it from us. You know, though, we're part of the open- source community. We believe in the open source culture and ethos, those sorts of things. So, now that things are being a little constricted and, once again, there's new platforms coming up I don't know how many micro blogging platforms- have come up in the past two years.

Ari: 12:29

It's just, it's kind of crazy. The social space is becoming more difficult and more hard to access data, right. So, and because of privacy concerns and a variety of other things, so Web3. Now we have a ton of data that's available on wallet interactions, on contract IDs, on transaction levels, those sorts of things. So a lot of that information can be surfaced and utilized from a marketing perspective to understand consumer behavior. It's kind of cool. But once again, we get to this point in time where you have a lot of data, right. So, let's say, you want to combine excuse me social data and Web3 data, right, yeah, blockchain data is structured, right, serviceable Social data is not.

Ari: 13:21

It is unstructured. It's full of let's face it miscontext, misdirection, misinformation, duplicity, jargon, all that kind of stuff. So you have to figure out a structure around it in order to create usability to it. This is where AI comes in, right, looking at this mechanism of normalization of data and going through the data to understand. You know, how do we look at users holistically across multiple different platforms and different data forms, but also how do we cluster different audience segments together to look at different attributes and different associations. And this is exactly what we did with a couple of different brands already, right, and it's not AI. So I have a hard time. And once again, gardening and Mellon University is one of the birthplaces of AI, right, because you know, herb, simon and Newell are one of the folks at Coindeterre and we have programs on AI. We've had one of the first undergraduate programs actually on AI in the world, and so the aspect of AI that you know, we do quite a bit and we've had for a long period of time.

Ari: 14:34

Is machine learning Right? And we do machine learning on just about anything and everything these days? So we're doing levels of AI which are very task-oriented. So that's currently being done. But I think with all of these different technology phases Web 3, you mentioned metaverse amount of data that you can garner from a head headset and somebody interacting in the metaverse explodes. The amount of data you can collect, right, and once again, this is you know you're wearing a headset, so I'm collecting a whole bunch of biometric information around eye detection, eye gazing, temperature, whatever, right, and you can imagine a brand. Who's really interested in this? Right is this? Is this product boring you?

Ari: 15:22

or is this product exciting you if you can imagine somebody like you know BMW or Porsche putting somebody in a virtual driving seat, creating haptic feedback on some kind of device and allowing them to do an experience that they might have at a dealership driving a car in a virtual setting. Right, they didn't want to monitor all kinds of things, right, and it also might be a way to customize the experience once you actually get your new BMW i4, i5 or i6, whatever that, whatever I version they might have, right, because now I know a little bit about your perception, your field of depth, your needs, your custom, your comfort level, all that kind of stuff right. So the mechanism associated with customization, but it's predicated on collecting a lot of data and analyzing them right, it's interesting.

Joeri: 16:21

You know I come from like every world probably in the marketing come from the Web2 world with social media and stuff. With good analytics, you don't have the data anymore than you had before, with all you know those blockers and stuff and the fact, the privacy issues, and you know, also, with email marketing you are not sure about the data anymore because it's not, you know, measured. Everything is not measured as it used to be. But in Web 3 you have so much more data it feels like now, listening to you so much more opportunities. Of course, everyone is not there yet. You know, because they need to have this headset, they need to be able, you know, to be uh, would say uh, to be measured, to be the data should be captured. I'm curious to hear you know, how should we now, as a marketer, adapt because we are shifting and you know Web 3 is coming. It's not there yet, you know. Do we need to anticipate, take certain actions? Should we adapt our marketing already with everything that is changing, even if we are not there yet?

Ari: 17:23

yeah, I think. I think the key is get ready right. A variety of things took marketers kind of by storm right and cause a bit of disruption. First the internet and then sort of web 2 and social right and you know, large social platforms sort of pushed around marketers right, changing their algorithm, even though their life blood, right, all the revenue is predicated on advertising marketing. They still were like, oh, but we're here for the consumer, which I understand, the consumer is a major stakeholder. But they're doing capital, you know, surveillance capital, using the consumer almost as a product associated with grabbing their data to provide to people who are actually, you know, paying the bills to keep the lights on right to expand the company which are the advertisers

Ari: 18:19

and the brand right. That dynamic needs to be more balanced, especially in the Web 3 world, right. But marketers also have to understand that infrastructure needs to be there, associated with support, right. So a lot of experimentation is going on, whether it's Nike or, you know, estee Lauder, a variety of different folks who are dabbling in sort of the web 3 space and looking at different mechanisms of creating utility, yeah, and experience management when it comes to web 3. But we also have to understand that there's regulations, right, there's policy implications and there's data collection issues, right. So I think you know, as we mentioned before, this relationship with IT is really important. Doing these experiments to understand what infrastructure is needed, how do you analyze and support the data right. So in the future, we're going to see CRM systems that are designed for web 3 specifically designed for web 3.

Ari: 19:23

Right, we're going to see all of the support infrastructure, whether it's sort of everything from insurance to security services to consulting. Right to even understanding ways of interacting with universities like myself. Right to experiment with the model and understand the value as the industry matures and comes into fruition. Right for the mass market, not for the early pioneers and the expert.

Joeri: 19:54

Yeah, yeah, you mentioned Web3 CRM, but I know already people in your business are building this stuff. They're not yet ready, but they are preparing of course. Then they need to go to the market and they have these challenges, of course, and then they come to me to ask my help. You mentioned a few words, Ari. I think experiences, experimental, experimental learning, is you know something for your social or your measuring social class? I should say that it's called a collaborated with brands like Netflix and stuff. So I'm curious to hear you know what are the insights you can share. Yeah, from analyzing social data and stuff with your students.

Ari: 20:36

Yeah, so I've had this class running at Carnegie Mellon for 13 years. It's a lot of fun.

Ari: 20:42

This semester you know, it's just a bunch of brands coming in like PayPal and Harley Davidson and Nike and HP and a like a variety of other folks that want to work with the students, because the students provide a unique assessment of the space, right, they're mostly millennials and Gen Z, right, and they have a unique outlook associated with perception of brands and how to utilize social data to understand, assess, insights and recommendations. So we call it experiential learning, action- based learning, project- based learning, all these different acronyms out there but it's being readily adopted across a variety of different higher ed institutions because there's a tremendous amount of value in it. Yeah, we could have a company come in, we could develop a theoretical model for them, but that would be, that would pretty much go on the shelf. I doubt it'd be able to be applied. My break on, you know, adoption, those sorts of things. So rather we become almost like a consultative sort of unit within the organization to understand needs, to understand sort of feasibility and implementability and scalability associated with solutions. A lot of the time we build experimental models for them to utilize around AB testing and a variety of other things, but the idea is really getting the students ready for things that are very much applied in the real world, right?

Ari: 22:19

So if you're in the social space, you have to get ready for everything, and we talked about brand hijacking, we talked about misinformation. We talked about, you know, audiences, misconstruing content, trying to cancel brands and deep- faking all these sorts of things that are going on. You have to be very much aware of that, as well as all of the different behavioral and normative dynamics within a social platform, right? Whether you're an X meta, Snapchat, macedon, post, spill, whatever one you want, it becomes really confusing from a market error's perspective. Which one do you invest time and energy in, right? So you have to do some experiments to understand audience uptake and value for the dollars spent and the resources spent in terms of understanding and assessing and being a part of the platform. We haven't even talked about Discord, right?

Ari: 23:24

Yeah, server within there. So there's a lot that needs to be done and I think you can take this idea of experiential learning and apply it to organizations. So organizations are coming to CMU and saying, hey, can we create a mechanism where we could develop workshops, we could develop design thinking sessions with you to do proof of concept experiments, right? So it's like experiential learning in an organization, leveraging some of the experience and know- how that an institution like Carnegie Mellon can bring into the picture. I think that's really important because sometimes organizations really need to get out of their sort of organizational memory right. Get out of sort of their head and into a different way of sort of thinking about a problem statement, a problem space, but also understanding that there are constraints that every organization has right. So that's where a partnership really comes in and really, when it's done well, you can see value associated with it right away. Great.

Joeri: 24:33

I think that's a great message and advice, you know, to end this podcast episode with Ari. If people want to know more, but I know you can go and talk for ages, you know about the subjects if they want to find out more about you, want to connect with you or is there any place you would like to send them?

Ari: 24:52

Yeah, you can. Just, you know, write me a physical mail, right, and send it in an envelope. Yeah, so I'm on all the social channels? Yeah, I mean just you know you can Google me, go to my website, drop me an email, drop me a post, drop me a snap, snap, your reel. Whatever you, whatever way you want to connect, I'm happy to connect with you.

Joeri: 25:17

Yeah, also, I'm on all these channels, but I guess, yeah, there are so many notifications also every day you mentioned it so many new channels. Thanks so much, Ari. It was a pleasure to have you on the show, yeah.

Ari: 25:30

I really enjoyed chatting with you.

Joeri: 25:32

So, guys, what an amazing episode with all this value. If you liked it and if you think, yeah, this is really useful for people around you, be sure to share this podcast episode with them. If you're not yet following our show, this is a really good moment to do that and, of course, I would like to see you back next time. Take care, bye. Web3 can take your biz to new heights and you're ready to harness its power, but feeling lost and overwhelmed. Therefore, join my W3X Web3 mastermind. Send me a personal message for more info. You can find me everywhere on social media. There's only one person with my name, Joeri Billast. Talk soon.