Get Started >

257: Leverage AI-Generated Personas with Cary Lawrence

Cary Lawrence, CEO of Decile, is driven by a passion for decoding human behavior to help businesses market to AI-generated avatars, build loyalty, and drive growth through data.

We discuss Cary’s four-step framework for customer engagement: collecting first-party data, identifying high-value customers, tailoring customer journeys, and tracking behaviors over time. She explains how Decile leverages AI to create detailed marketing personas, allowing brands to engage customers more effectively and optimize marketing spend. She emphasizes the importance of understanding customer nuances to build genuine connections, enabling brands to achieve sustainable growth by aligning with their customers’ unique needs and preferences.

—

Listen to the podcast here

 

Leverage AI-Generated Personas with Cary Lawrence

Good day, dear listeners, Steve Preda here with the Management Blueprint podcast. And today, my guest is Cary Lawrence, the CEO of Decile, an e-commerce analytics platform that delivers business intelligence and customer insights to help its partners build loyal customers, optimize products, and drive profitable growth. Cary, welcome to the show.

Thank you, Steve. I’m excited to chat with you this morning.

All right, well, I’m excited as well. Let’s start with the first question, which I like to ask people because I like companies that are mission-driven. And I know that yours is one like that too. So what is your personal “Why” and how do you manifest it in your business?

Yeah, and I think when I think about my personal ”Why,” for me, it’s really about understanding human behaviors and what makes people tick at their core. How do you tailor a message for someone, whether it be your spouse or your child or your colleague or your client, to try to create kind of the most productive relationship for both. If I think about understanding what makes people tick and those human behaviors in my personal life, I think about my two small children, for instance. I have a 10-year-old daughter, Spencer, and not just recently turned 8-year-old son, Nash. What I learned from being a parent is that the way that my daughter needs to communicate and the way she learns is very different from the way my son does. So she’s very much a rule follower and wants to do everything that a coach would tell you, whereas my son learns by doing. Similarly, when I think about my husband, one funny story that I had to learn along the way was we both lead our businesses, and he was venting one day about some of the challenges he was dealing with, with his architecture firm and clients and the struggles that we all encounter. My immediate impulse was to just solve the problem and tell him what to do. Finally, he had to say, time out. He’s like, what I want is just for you to listen and let me vent. I don’t really need a solution right now. Once I understood that, I was able to be a much more supportive spouse in the way that he needed me in that moment. But I think to your question about how does this manifest in my work at Decile, it’s very similar. I think that taking that sort of understanding of human behaviors, I’ve sort of translated that into trying to understand consumer behaviors online. And as we increasingly spend more time online, like what are those behaviors? And even more specifically, how do marketers best lean into understanding those different human nuances and understanding that not all of their customers are the same. There’s going to be different variations of customers. And so really understanding how to talk to someone, where to find them, how to really think about how to conduct that consumer journey with them over time. So anyway, all about understanding those different people nuances.

Yeah, that’s really interesting, and especially how you can do that while just clicking around on the screen and by, I guess, based on what they select and what they respond to and how they respond and what they follow. I wonder how that even works. What are the ways that you actually can monitor and discern these behaviors?

Yeah. Well, for us, it’s not so much about looking at all their click behavior online. It’s really all comes down to that first-party customer data. So I see the marketing world is moving towards more, what we would call identity-based marketing. So there’s definitely a need in today’s environment for marketers to start to collect that first-party customer data. So when we think about the management framework and how we think about optimizing the ability for our clients to help them grow profitably, it’s really kind of a four-step process. And so the first step is literally you need to start collecting that first-party customer data. And by that, I mean you need something that identifies you as a unique individual. So an email address, phone number, home address. And what I would say is like as long as you’re providing an equal value exchange with your customers or prospective customers, they’re willing to kind of trade that information. So maybe it means they get a special discount or a special gift with purchase or what have you. So step one is literally just starting to collect that data. And then I think step two is really understanding again that not all customers are created equal and the world has changed a lot where you can’t just do what I think a lot of direct to consumer brands did 15, 20 years ago, which was, I just want to grow as many customers as possible and grow my revenue in that top line view. And that’s like what’s happened since is a lot of these businesses, their P&Ls and their businesses have been turned upside down because what they have been doing is spending a lot of money to just acquire any old customer and they’re not valuable customers. When we think about step two is really how do you identify those high value customers? By that I mean, who’s going to be a repeat purchaser? Who’s going to sign up to be a subscriber. And honestly, it doesn’t really matter as much how much it costs you to acquire that customer if you know that they are gonna be high value. So being able to predict in the future, this person looks like they share a lot of the attributes of my high value customers. They’re gonna have a larger basket size. They’re gonna kind of continue to purchase from me. They’re gonna sign up to be a subscriber. So you really need to be able to think about what we would call cohorting those different customers. So understanding by any dimension, not just who’s likely to be high value and a loyal customer, but what’s the difference between how males versus females buy or what’s the difference between someone in a suburban environment versus an urban environment? Really understanding those nuances is important. And then step three is, okay, now you’ve acquired your customer, you need to start that conversation with them immediately. You need to understand as soon as you acquire them, which kind of persona do they look like? Are they like a Cary, an urban female mom who really likes to travel and go to fine dining restaurants? Or maybe they’re more of a young tech innovator just out of college. Who knows what the different personas are, but you have to understand not just who they are, what are those purchase metrics, but also the softer things like demographics, psychographics, behaviors. And then you have to kind of understand, okay, well, here’s where I was able to find where they’re spending time online. They were on Google or they were on Meta. And here’s the type of products that, I think, I see that they have an affinity for. And here’s how many days it seems like they’re going to need before they purchase their next thing. So you want to be immediately understand as much as you can at the beginning to kind of continue to nurture them. And then the final kind of step in the framework that we think about is like, okay, we’ve got them, we understand who they are. We need to track those behaviors because people change, right? Like maybe I’m not always going to be a high value customer. Maybe I’ve lost my job and I’m no longer buying my expensive skin creams every other month. Maybe that’s changed. Or maybe a marketer has shifted what products they’re offering and they found that that actually brings in a different type of customer. We’ve seen that happen with some clients where, one example is it was a skin care company and they traditionally sold products to men. That was kind of their thing. And then they started to, they wanted to widen their customer pool. So they developed some products specifically for females. But what they did is they started immediately shifting all that budget into marketing towards females. They kind of forgot about their core customer, that kind of initial male customer. And then when we looked at the analysis within Decile, we said, you know what, those males are actually on average a $10 higher lifetime value than the female. So yes, let’s launch the new product lines and attract new customers, but let’s not forget those existing customers too that drive a lot of that value. So that’s kind of the final piece is being able to track those behaviors over time and make sure you’re continuing to refine, to kind of bring in those high value customers.

Yeah, that’s a really big challenge to be able to do that, especially in a granular way because you can put people in three or four brackets, but we are a lot more different than just being a suburban mom or being an empty nester white male. I mean, there are so many people, so many different interests and affiliations and so on. So how many different buckets do you put people in? I mean, what are the variations? When you do the market to these people, how granular can you get?

Very granular. There’s a couple of things that we would actually have marketers start with first. And the first, and this is kind of the history of our name, a decile, if you think about it, it’s when you divide something into 10 parts. So when we talk about your top decile, that would be your top 10% most valuable customers in that context. So one of the first things we’ll do is actually look at a client’s customers by decile and see, okay, who’s really driving their profitability? And oftentimes it’s that top 20 to 25% that are driving 80% of the revenue and profit. So that alone can be eye-opening for marketers. Like, oh my gosh, I don’t want to waste all my marketing dollars targeting these lower value customers when 80% of my business is really driven by this top 20%. So that would be kind of an initial analysis is to look at it by those deciles. And then another analysis is, to your point earlier, it’s like, okay, well, we need to understand more about who these people are and what their interests are, not just their purchase behavior. So that’s where we have AI-generated personas, that the machines are basically doing the work to pull in all that purchase behavior, but also layer that with hundreds of different attributes. And that’s things like, where do they live? What’s their demographic? How many children do they have? What are their buying behaviors? Which brands do they tend to have an affinity for? And so, then for each brand, it’s unique. It’s not always the same, Steve. So like if you’re Brooklyn and you might have very different personas than if you are Athletic Greens, they’re like very different types of customers. So it’s unique to the brand. But if you’re just starting out, you could even start simply from a segmentation standpoint, think about it like, okay, I want to segment out, let’s say my males from my females, or I want to segment out people who buy this product versus that product, or I want to look at my high value customers versus my low value customers. And then the nice thing is once you’ve created these segments or cohorts, you can then activate them on your advertising platforms and remarketing platforms. So you can say, okay, I want to create an audience of just my, let’s say top 10% highest value customers. And I’m then going to send them as an audience into meta and let meta do a lookalike to find people just like them. So what you’re basically allowing the marketer to do is be very efficient with their marketing spend because they’ve already sent a really strong signal into the platform to say, these are the people that I know will drive my business. I want more of these people. And then the advertising platforms will be able to do their thing to help find more of those folks.

Yeah, I like this idea of the AI-generated personas. And then you focus on your, it’s like the Pareto principle, the 20% of your prospects that will generate 80% of the result.

That’s absolutely right. Yeah, it holds true.

10% is even more refined. So how many different profiles can you develop? Let’s say you take your top 10 percent, and then you have all these different personas there. So how many different, with AI, marketing campaigns would you be running for this 10 percent depending on the different personas that they may have?

It’s really up to the individual marketer. So let’s say they just want to take their whole customer base and see how that breaks into personas. The data science will do its thing and they might show, okay, there’s five distinct personas, let’s say, of types of people. But then we have some clients who want to take it one step further and they might say, okay, I know I want to just focus on my top 10%, my highest value customers, but tell me more about the differences between those. So you could again run a persona set on just that slice of your customers. And then they have to think about, okay, well, how much marketing budget do I have? Can I apply? And what are my goals? Do I want to do a big product launch? If so, let’s say I’m launching, just to stay in the skincare example, let’s say I’m launching a new skin cream. Well, I should find out from my customers who’s most likely to like skin creams. Is it the older set? Is it the younger set? You know, female, male. And so by understanding who in their current customer base tends to like those products, then they can be really smart about who they’re targeting online. But it really does, there’s a lot of kind of nuance based on that individual brand. The key, again, it all comes back to like having that data at your access and being able to understand your customers and what makes them unique, what makes them different, and what do they respond to? Some people might only want to spend their time on TikTok. Some might only spend their time on Google. Some might say, I only want you to send me an email every two months, don’t email me every day. I’m just gonna unsubscribe and say stop if you’re too much. Or one thing I will often tell marketers is think about the channel and how you’re communicating. I would say about two years ago, the SMS or text messaging for marketing became really popular. That’s a very intimate channel. You don’t want marketers blasting your cell phone, texting you all day long with just a blanket offer that’s like, hey, think about X beauty brand or we’re having a sale. You’re like, okay, great. Well, stop bugging me. But if they said, Cary, we know that you really like the skin serum and you typically buy it every… This is kind of in their head. You buy this every two months and you only buy it when it’s on sale. Here’s a 20% gift card. This is on sale today for you, Cary, and I’m going to give you a free sample too because you’re a very loyal customer. Like those kind of things would make me say, oh, okay, it’s time to go buy that skin serum.

Okay. So that’s interesting. So one-on-one marketing is getting more and more refined. You basically target just the right prospects for you, and then you segment them, and then you personalize the AI. What are any other trends? So if you look forward into the future, the next couple, how do you think this industry is going to evolve? What are the new things? What are the boundaries of technology and approaches that are being experimented with right now?

Well, AI is definitely the game changer. Probably no big surprise there. And I just was at a conference on a panel with a group talking about all the impacts of what they expect to see and how the machines are going to change things. So I think a couple of things. Like the first sort of big change came in the form of content. So using AI to help develop creative iterations and content. And so now, whereas it used to be a little bit more time consuming to develop that content and it was too expensive to customize it for every single customer.

Now, the machines and AIs made that a lot easier and more accessible. So they can very quickly find out, okay, this is a creative that will match with this customer and here's this one on a different background than this one. Share on X

And so like the content development or I want to create an ad focused on this new product that will make people feel inspired. Like it’s just gotten a lot easier on that front. Now, the other, now the evolution, obviously I talked about using AI in the context of personas and audiences and targeting, but you now start to hear about marketers using it for things like customer service. So a lot of people, like if you go to their website and you talk to their chat person, it’s not just going to be a generic, how can I help you? What’s your order number? They’re going to know about you. They’re going to be able to offer you more customized, personalized feedback. Another interesting use case was thank you notes, believe it or not. So some brands were using kind of AI to generate personalized thank you notes to send to customers as part of, I know it’s mind boggling. And then the newest thing that is kind of blowing my mind is a lot of marketers think this is the holiday season where you’re going to start to see shopping agents really start to take off. By that I mean, there are machines that will do your shopping for you online. They will say, okay, I might say, find me the biggest discount for this type of clothing or find me this color sweater. The machines start to go and search everywhere and find those options. So the question that marketers have is, well, how do we start to market to a machine instead of a person? And so it creates a whole bunch of kind of interesting, new potential takeaways. And then there’s this idea of, you have to still maintain that brand authenticity, right? At a certain point, some people don’t wanna feel like they’re talking to a machine. They wanna feel like your brand is authentic and it's not just scraping the web for anything. Share on X So there’s a lot of big questions that marketers are kind of grappling with now, but definitely, AI is the big game changer right now.

Very interesting. So we are also experimenting with AI and we have ran into a problem that I write a weekly blog, which I send out to my email subscribers, and then we want to turn it into a website blog, which will have the right keywords and so we started using AI to rewrite the blogs with keywords, and it essentially made it a very sterile and it filtered out the humor from it and the personality and it became so bland that I don’t want to publish it. It just looks old.

Yeah. No, I mean, that’s absolutely right. You want to maintain your own personality and your brand. It’s both interesting and a little bit scary because they are going to get better and better. Some of the things that I think about are like, my voice is out there on the web from different marketing panels or interviews. What happens when the machines can grab your voice and your type content and then people get confused? Is this really Cary or is this just a machine version? So there’s a lot that I think we’re going to be thinking about. I would say for now, though, the main thing that companies and marketers should start thinking about is just this idea of personalization. So like, let’s use the machines to kind of help you do more personalized marketing outreach, because there’s going to come a time in the very near future where it’s not going to be enough to just have your site-wide discount for everyone. That’s also a waste of money. I tell marketers that all the time. I’m like, let’s not give the same 30% discount to everyone. You have people who will buy full price every single time. Save the big discounts for those discount-sensitive shoppers and target them with ads specifically so you don’t have to give the same discount to everyone. Similar for like a free shipping. Sometimes you can use data science and your data to understand, okay, do I need to give free shipping for just a $50 purchase or can I raise that to $100 and still get the same behavior? So, again, it’s understanding the consumer behavior. What’s that threshold where they’re going to say, that’s too much. I’m like I’m not going to put that much in my basket, or I will add that extra shirt to get free shipping. Like, marketers are going to need to know all those nuances to be successful.

Yeah, that’s definitely happening. I ordered some print materials and they had the minimum $125. I was thinking, okay, what else can I order to reach the $125? And my order doubled just because of that. I brought forward something that I wouldn’t have ordered at this point.

Yeah, they understand you. They know what they need to do to push you into that next version.

It could make me a bigger ticket customer.

Yeah, exactly.

Yeah, that’s cool. So before we wrap up, Cary, I’d like to ask you, what is it that you’re working on that most excites you right now?

Yeah, I think for one of the big projects that Decile is working on right now is something that we would call Anomaly Detection. And that sounds really fancy and confusing, but what it really means is helping a marketer understand what in their business has changed recently. So that might be, what are the consumer behaviors that have changed? Which are like, is there like repeat purchase rate all of a sudden spiking, or did  they all of a sudden get like an infusion of a whole bunch of new customers because they had some viral moment? So we want to be able to like have their data surface these alerts to them so that they can understand something has changed in either my customers’ behavior or something in my business. Maybe all of a sudden, like I’ve lost 10% of my subscribers. Why did that happen? And so being able to like lean into not just here’s all your data, but like how can your data help you run your business in a more meaningful way so you know exactly when something important has changed.

Yeah, that’s huge. When you can have those little triggers. I think LinkedIn does the same thing that when someone changes jobs, then if you’re on sales, you’re going to get a prompt and then you can congratulate them or do something, use that excuse to personalize the message to them.

That’s exactly right. That’s what you want. You want to know when something’s changed and how you need to react to it quickly. But I think that will be an exciting moment for Decile, but also just for marketers in general.

Okay. All right. I have one more question I’d like to ask you, and we haven’t spoken about this, but as a CEO, what do you feel is the most important question for a CEO to ask themselves?

I think that the most important question that I will often ask myself is, what is the most important thing that I should be spending my time on today to generate the atmosphere that I want my team to have? So that might be, and again, as a CEO, you’re wearing many hats. So it might be today, there’s nothing more important than me chatting with Steve and letting everyone know about what’s important about Decile. Or it might be today, what I really need to do is to work with the product team to understand how we can meet this deadline? So I think it’s really understanding what is the most important part of your day that can really drive the needle. Because you could definitely have just sort of a reflex to just spend all your time in spreadsheets and just be thinking about everything in a bit. But I think it's really being able to focus on what really is the thing that you can do to move the needle. Share on X

Yeah, I love that. This is great because we all have to-do lists with a million items on them, but are they going to all move the needle? Probably not. Much of it is busy work. So how do we pick some big thing? And if we pull that lever, we’re going to be fine. Even if you miss a couple of other things from our list.

That’s right.

That’s great. I love that. All right. So Cary, if someone would like to learn more about what Decile does and how you guys help e-commerce companies to get closer to the customers, to make them more loyal, to find the top 10%, to drive revenue, where should they go and where can they connect with you personally?

Yes, thanks for asking. Decile.com is our website address. You can find me on LinkedIn, Cary Lawrence at Decile. And my personal email is cary@decile.com. So please don’t hesitate to reach out.

Okay, that’s very simple. So do reach out to Cary Lawrence, the CEO of Decile. If you want to build a better relationship with your customers, you’re going to drive revenue, you want to be at the forefront of AI-generated customer, what do you call it, not profiles, but the…

Marketing personas, yeah.

Personas and avatars, I guess. And if you enjoyed this episode, then do follow us on YouTube and follow our LinkedIn channel as well, Steve Preda Business Growth, and you will find exciting new entrepreneurs twice a week coming to our show. So thanks for coming, Cary, and thanks for listening. 

 

Important Links:

Summit OS | Stevepreda.com

PO Box 400, Rockville, VA 23146
Phone: +1.804.332.1307