The CMO Whisperer, hosted by former Forbes writer Steve Olenski, is a curated executive platform and marketing leadership podcast where CMOs, brand leaders, and senior operators share candid insights on leadership, brand strategy, and growth. Spanning business, sports, and media, it goes beyond marketing to reveal how top leaders make decisions and perform at the highest levels.
Hey. Welcome to The CMO Whisperer Show. I'm your host, Steve Olensky. Part marketing practitioner, part ad agency veteran, part journalist, I was a writer for Forbes for ten years. I've had so many insightful conversations over the years with business leaders to athletes to celebrities to, of course, CMOs.
Speaker 1:The only difference now is instead of sharing those insights through written form, I'm doing it this way. Today's episode comes to you live from Possible where the badges are swinging, the buzzwords are flying, and somewhere nearby, someone is promising to reinvent marketing before lunch if that's possible. Today, I'm joined by Ron Cohen, SVP of practice leadership at Claritas, a company operating at the intersection of data, identity, and AI, helping brands understand who their customers really are and how to reach them with more precision, relevance, and result. Now Ron spends his time focused on one of the biggest challenges in modern marketing, turning scattered signals into smart action. In a world full of noise, he works on helping brands make better decisions about audiences, personalization, targeting, and growth.
Speaker 1:Ron Cohen, welcome to the CMO Whisperer podcast. Well, let's just go. Right? Let's get over the amenities and our favorite colors, and let's just move in. Talk to us about you, what you do first.
Speaker 1:What lane do you live in every day, and what problems are marketers bringing to you the most right now?
Speaker 2:I have a lot of hats. Most of my focus is on identity, identity resolution, personification, audience creation, and pain execution, and taking the signals you get and refining, demising, but do it again. Only better.
Speaker 1:So is that... That's a con... That's kind of a consolidated view of what marketers are coming to you with. Right?
Speaker 2:Yes.
Speaker 1:Give us a little step below, a little granularity. Does it depend on the vertical? Does it... Like A
Speaker 2:little bit. But if you start with, we resolve the identities of existing customers and potential customers who come to the websites of our clients and don't self identify. Who are they? What pages are they looking at? Is there a difference between them and the general website visitor or the general population?
Speaker 2:Which way does
Speaker 1:this skew? Right.
Speaker 2:Who should they be targeting? What is... What's that segment look like?
Speaker 1:So we hear a lot about data, but not enough about identity. Why is identity such a critical critical issue for marketers today?
Speaker 2:What could... And how you can reach them with what offers. That's all identity based. We have a ton of data, but if you can't resolve the identity of your clients, customers, or prospects, how are you gonna action? Act on that data.
Speaker 2:How are you gonna put it into action? It's like insights without actions. Yeah. Actions without insights.
Speaker 1:Yeah. Those who listen to my show and those who know me, it's... I'm a very simplistic guy. Just like I'm listening to you going, yeah. It really is that simple.
Speaker 2:It really is
Speaker 1:that simple. Is is that something that, you know, brands overthink, overcomplicate?
Speaker 2:Like They don't wanna necessarily take the time or spend the get that level of detail. They think, I'll just advertise, and people will buy. If you're McDonald's, maybe that works. It does work if you're McDonald's. Right.
Speaker 2:Most people aren't McDonald's. Most brands are not McDonald's. Need deeper insights into your prospects, into your customers, and into what makes them tick. They're gonna respond to, and in what channels they're gonna respond, and in what tone you need to communicate. I'm seeing what language you need to communicate if your message is gonna resonate with them and get them to take that next step, whether it's click or buy or
Speaker 1:Whatever the case.
Speaker 2:Whatever it is.
Speaker 1:Yeah. So my guess is there's a lot of brands, and we're not gonna name names, but who do it without those extra granularity steps for lack of a much better term. And then look at their results and look at their ads and say, oh, my advertising is not working. Well, it's not
Speaker 2:That's true.
Speaker 1:It's not that simple, though.
Speaker 2:No. Usually not.
Speaker 1:Right.
Speaker 2:Message. AB testing. Okay. Is image a better than image b?
Speaker 1:Right.
Speaker 2:Maybe. We talk now about a to z testing because image a is better for some people, image q is better for other Image c is gonna be the best for yet other people. It isn't an a or b.
Speaker 1:I like that.
Speaker 2:It's a to z and every letter is good for something. That's what our optimization technology is about. That's what... That's how we leverage our data. We have so many segments, we have so much data.
Speaker 2:Something is right for each type.
Speaker 1:I love that a b... Like, don't stop at b.
Speaker 2:No. Don't stop at b because you'll never get the right answer for the majority of the people you're trying to communicate with.
Speaker 1:I love that. I absolutely love that. So if if brands don't have a clear view of who their customers across channels and devices, say, what what breaks first, if you will, without that clear view?
Speaker 2:Probably their budget.
Speaker 1:Okay. That... What comes next? Is it the relationship? The relationship or some
Speaker 2:smart competitor messaging at the right time and the right channel, etcetera, etcetera. There's so much competition, and someone's gonna do it right. And if you're not paying attention, they're they're probably gonna steal your customers.
Speaker 1:Even at McDonald's.
Speaker 2:Even at McDonald's. Absolutely. One slip, we won't name any names. No. No.
Speaker 2:No. You don't have to look too far back.
Speaker 1:Nope. One slip Nope.
Speaker 2:That comes from know know your customer.
Speaker 1:Exactly. It's not a to... It's a to z. So how should marketers think differently about audience creation today versus the old, say, static segment? There's nothing wrong
Speaker 2:with the static segment model as a starting point. Right? That's our playbook. Let's code your customer file. Let's resolve the identities of your website visitors, and let's start with static segments, and then start testing a to z.
Speaker 2:What works for segment one? What works for segment two? What's... It's creative, it's channel, it's audience. You have to optimize everything all the time.
Speaker 2:We have new branding now, new shirts out this week that says optimize everything all the time. It's a lofty goal, but it's the right idea. It's exactly how you should be thinking, especially with AI because it can think out thick people. It's not that it's so smart, it's that it computes things faster. Mhmm.
Speaker 2:Machine learning really, I'm talking about now. You can optimize everything all the time or at least you can have a goal of doing that.
Speaker 1:Well, yeah, like shoot for the moon. What the hell?
Speaker 2:Exactly. So if you start with static segments, I've already put you ahead of half the class. Because those segments are built based on real observations, real research. And then you start optimizing for whatever it is you're trying to sell, service product, whatever, in different channels, with different messages, to different segments. And each time you learn something, or the machine learning kicks in, and it gets better and better and better.
Speaker 2:And pretty soon, you know, your your return on ad spend is starting to look like, this is a no brainer. Just leave this on all the time.
Speaker 1:But you said something so important, you gotta stick with it.
Speaker 2:Yeah. You have to stick with it. The longer you do it, if you're doing it right,
Speaker 1:the
Speaker 2:better it gets.
Speaker 1:The
Speaker 2:harder it gets. Right. The better the results get. Optimize everything all the time, and you're never done. That's never done.
Speaker 1:Yeah. You don't optimize first quarter
Speaker 2:and stop. And then, yeah, let
Speaker 1:it ride. No. I love that. What's the biggest misconception brands still have about personalization?
Speaker 2:That's a good question. Some brands think they can't do it at all. It's not worth doing. They want mass appeal. If I do too niche, I'll limit my audience.
Speaker 2:I'll limit my sales. You can do mass personalization, mass optimization, you know, one little niche segment at a time. That's where you have to open your mind to the possibility that there are many, many, many niches, and each one can be messaged in the optimal way.
Speaker 1:Would the would the... Sorry to interrupt you. Would the misconception be that the brains think they can't personalize at scale?
Speaker 2:Or they get hung up on this one to one thing. Yeah. One to one.
Speaker 1:Sure. That was huge.
Speaker 2:Yeah. It was huge. But It's 300,000,000 segments. As a concept, it was a good idea. Let's, you know, focus on one person at a time.
Speaker 1:Mhmm.
Speaker 2:I would say focus on one type at a time. You can't have three hundred million seconds. You can't have 300,000,000 creatives. You can't track 300,000,000 messaging strategies.
Speaker 1:No. Unless you have a lot of free time. No.
Speaker 2:And even
Speaker 1:I know. I was just gonna say... Exactly. I like that. This question, I cannot wait because this word is so, pun intended, relevant to me, and it's the word relevance.
Speaker 1:Right? And I wanna get your thoughts on when done right, what does relevance actually look like to the customer?
Speaker 2:Question a lot. We do a lot of multicultural research and advertising. And you may ask, how important is it to you to see advertisements that have people who look like... Forget about the message, just the people. You see?
Speaker 2:If they don't look like me, don't pay attention. And if I am paying attention, I don't believe them. Or what they're saying isn't relevant because they're not like me. Yep. It might be relevant to them, but not to me.
Speaker 2:So it starts with that. It starts with can I relate to whoever the messenger is? And then content, imagery, music, audio, language, themes. Now I'm thinking multicultural, but it's... It applies Yeah.
Speaker 2:Beyond that. It's more obvious when you're talking multicultural.
Speaker 1:Yes. It's more glaring.
Speaker 2:Yeah. When you get it wrong,
Speaker 1:it's more obvious. That's what I meant.
Speaker 2:But it's really no different for the general market or any segment. It has to be code.
Speaker 1:It has
Speaker 2:to be the messenger and the message and everything else. The language, the sound, the feeling.
Speaker 1:Everything. The emotions. We did convenience store research, and
Speaker 2:it turns out the smell is more important than anything. Wow. The smell for certain types.
Speaker 1:Yeah. Of course. Yeah. That's interesting. With the lighting, with the music.
Speaker 1:Right.
Speaker 2:With the shelf management, the first thing that hits you when you walk in the door is the smell. Critically important. So we have all senses.
Speaker 1:Look. I... I'm a big believer that forever marketers have not given consumers enough credit. And what I mean by that in this context of consumers are savvy enough today to know that if you send me a message that has nothing to do with me, there's really no excuse to go, wait a minute. Hold...
Speaker 1:Especially if I'm an existing customer. Right? And we all do it. We've all gotten them and going, what? Like, not even so much you just tried to sell me the exact same thing I just bought, but you're sending me something that has nothing totally relevant to me.
Speaker 1:And I just think marketers don't or underestimate. I love your thoughts or not that you think brands just think consumers are, oh, they're good over it.
Speaker 2:Mostly, won't. Right. They'll think they've moved on, and they're not for me anymore.
Speaker 1:Alright. We gotta we gotta tackle the two letter word that's very pervasive in this conference and everywhere else, and that's AI. And I wanna start with... I wanna get your thoughts on not naming names, not naming tools, but kind of, like, where's the Ron Cohen of what's real and what's kind of snake snake oil or close to it?
Speaker 2:Okay. So let's separate AI from machine learning. Okay. Machine learning is real, and it enables all kinds of applications that are not possible without it. Too much math.
Speaker 2:Too many things to look at and analyze in real time to be useful. You're gonna optimize a campaign. It doesn't make much sense to make those decisions if the campaign is over.
Speaker 1:I wouldn't think so.
Speaker 2:Humans can't do it. It has to be machine learning. I think that's real. You look at data science and pattern recognition, machine learning is great for that. Some AI is very useful.
Speaker 2:If you design an agent to do something that's gonna save you hours per day, per week even, that's useful. But when you ask an AI a question at a client, I can't take credit for this, I love this answer. He said, it's kinda like ask the intern to do something. Maybe the intern does a good job. Ask the intern.
Speaker 2:And and you say, wow. We should hire this guy. But most of the time, he doesn't do such a good job. And you say, who the hell hired this guy? It's right.
Speaker 2:That's that's what you get when you talk to a a chat GPT or... I don't, you know, I don't mean to single that out. Any AI large language model is hit and miss. If you ask it certain kinds of questions and expect certain kinds of answers, it's like asking me. If you get lucky, probably not.
Speaker 1:It's an interesting analogy.
Speaker 2:It is. Right? You can't take credit for it.
Speaker 1:No. There's a lot of there's a lot of validity to... You just... I don't wanna see you put it out into the ether, but in a way, kind of like the intern analogy of, sure. I could ask an intern to do this.
Speaker 1:I don't know what. They may hit it out
Speaker 2:of the park. Right. Great. Probably not.
Speaker 1:Probably not. Interesting. So... Okay. I'm I'm a CMO.
Speaker 1:I'm a head of brand digital. I'm listening to this, and I'm going, okay. AI, I've I've learned enough that it could improve targeting and messaging, right, when done right. But I don't wanna lose that human side of market. How do I balance that?
Speaker 2:And with human supervision, we talk generative AI. We're gonna let the AI generate your creative and just serve it because it knows everything. Okay. That's gonna end badly. It's always going to end badly.
Speaker 2:You need human supervision. You need human training. You need human validation. If you're gonna use real AI, not not machine learning stuff, but real AI, you need expert humans training, prompting, validating, checking before it goes out the door. I don't know any other way to to answer that.
Speaker 2:I'm not a skeptic. I'm a I'm a heavy user of AI, which means I've had a lot of experience with unexpected results. But that's gonna happen. It's gonna happen.
Speaker 1:You have to know that going in.
Speaker 2:Yeah. And you'll have to plan for it.
Speaker 1:And you have to plan for it.
Speaker 2:Yeah. It's nothing... There's nothing new under the sun. It's like, if you don't make any mistakes, you don't learn anything. And if you're right every time and you win every time, what have you learned?
Speaker 2:Zero. Right. Expect to get unexpected results. Expect it to not do what you wanted it to do the first time. Probably not the second or third time either.
Speaker 2:Expect that. And that's okay. That's how it is. Right. It has to be okay.
Speaker 2:Right. Subtle term is stay away from it. Don't use it because there's no... It'll it'll always be right.
Speaker 1:Exactly. Exactly. Okay. Put your prognostication hat on. Where do you see audience targeting heading over the next few years?
Speaker 2:Machine learning. We know a lot about your existing customers. We know a lot. Even if nothing changes, which is never the case, we know who their lookalikes are. We know where you can find them and how you should message them.
Speaker 2:Alright. Let's say, hypothetically, we've got all the lookalikes, and we'd message them all, and now you need to grow up because it's all about growth. It has to be about growth. Who's next? And for that, you need transactional data.
Speaker 2:You need behavioral data. That's, I think, where where segmentation is going. That's where audiences are going. It's predictive based on real time data analysis, machine learning, and optimization. Again, trial and error.
Speaker 1:Alright. Last question. I'll get you out of here on this. If you could give every CMO listening one piece of advice about using their data better starting tomorrow, what would it be?
Speaker 2:Collect more and collect a lot of information. Mhmm. Retailer. What do you really know about your customers? Especially if they're not in your loyalty program.
Speaker 2:Even if they are, you don't know very much. It's worth investing in learning more about your customers. Or your prospects even. People come to your website. People who come to your store and don't buy anything.
Speaker 2:Pay cash. Well, that's not so much. You collect more. Know more about your customers. Know more about your shoppers.
Speaker 1:Yeah. Your window shoppers even.
Speaker 2:Whatever that means. Yeah. Know more. That's that's my advice.
Speaker 1:Okay.
Speaker 2:Starts with that.
Speaker 1:Gotcha. Well, Ron Cohen, SVP of practice leadership at Claritas. Thank you so much for your time. Appreciate
Speaker 2:it. My pleasure.
Speaker 1:Well, that wraps up another episode of The CMO Whisperer Show. I hope you shared this episode with your friends. And if you have not already, please subscribe to be kept up to date on all the latest episodes. And if you're so inclined, leave me a review on your favorite podcast platform. Thank you.