440. AI Meets Behavioral Economics: Predicting Consumer Behavior w/ Sam Albert of Behaviorally (Refreshed Episode)

Welcome to episode 440 of the Brainy Business Understanding the Psychology of why People Buy. Today's episode is a conversation with Sam Albert of Behaviorally Ready? Let's get started.

You are listening to the Brainy Business podcast where we dig into the psychology of why people buy and help you incorporate behavioral economics into your business, making it more brain friendly. Now, here's your host, Melina Palmer.

Hello. Hello everyone. My name is Melina Palmer and I want to welcome you to the Brainy Business Podcast. Today's episode is all about combining artificial intelligence and behavioral economics to create predictive models that businesses can really use to increase their sales and connections with customers. Today's episode is a refresh of my conversation with Sam Albert of Behaviorally, who will share about his own background and the work their team has done with Florida Flash AI, including sharing some great examples and case studies. Sam has over two decades of experience in the research industry and serves as Chief Digital Officer at Behaviorally. He is responsible for ensuring that the company is efficiently organized, aligned with the firm's behavioral framework, and that digital and physical operations support the overall business objectives consistent with the company vision. Previously, he held the role of Chief Operating Officer at PRS In Vivo Globally. After leading the quantitative research team in the Americas for eight years, he started his career as Research Director at PRS In Vivo's predecessor, Perception Research Services. His experience includes managing major client accounts, providing strategic direction on research methodology, project implementation, execution and interpretation of results. Now, don't forget, there are links for everything, including my top related past episodes and books waiting for you in the show Notes for this episode, which are found within the app you're listening to and@the brainybusiness.com 440 all right, let's talk about AI and behavioral economics. Sam Albert, welcome to the Brainy Business Podcast.

Thanks for having me.

Absolutely. Very excited to have you here. I always love to have our guests introduce themselves. So please tell the audience a little bit about you and your background and how you got into this crazy amazing world of behavioral economics and behavioral science.

Well, I guess may not be too unlike a lot of other people but I fell into it so wasn't on a charted path and kind of grew up as a kid saying I'm going to do B side and work in behavioral science when I grow up. Yeah, wasn't anything like that. So basically coming out of college said I was actually a pre law political science major, said well need to find something to do, we're going to go to law school, had all these dreams of doing all of that. And then this is going back always when they had actually newspaper job listings. Imagine that. Saw an internship for a local company and applied to it and it was for prs. And so ended up actually getting that job and doing some work on the internship side, which was mostly shopping and that kind of thing, but ended up getting a job on the research side after the end of the internship. And it kind of all just started from there. Started working on the client side, working on research projects. And back then, I mean I like to say we were doing essentially bsci and before it was a really B side, before it was called that we spent a lot of time looking at, not just measuring response, but also observing. So one thing that we did was we had tools and techniques even back then that were for the time very tech forward. So eye tracking. At the time there weren't too many people using eye tracking and that was one of our tools that we used to understand behavior versus asking people about what they did. It kind of started from way back then from those kind of basic tools to where we are now with Behaviorally and some of the tech forward tools that we're using today to really understand consumer behavior.

Yeah. And so for the background of listeners. So PRs changed their name to Behaviorally.

I believe in January of 2021.

Right, exactly. Jan. Feb. So we changed the name. The name, I think now number one is cooler for one to start with, but two is more apropos for what we're we're doing. Really trying to understand consumer behavior and more importantly help our clients essentially make good decisions based on that behavior. So it's on both fronts. Behaviorally really is apropos for what we do.

Yeah. And so then, so from a poli sci law interest to now being the chief digital officer at Behaviorally, was that an on the job training like you said? As you guys were new, you know, early on in a lot of the field, you know, you're kind of, there's a lot of that on the job insight that's so valuable, especially early on, versus you know, going back and doing academic research or anything like that. Like just a little bit more about you and your interests and how that's kind of shaped some of what's going on in your work.

Yeah, it's probably more of an on job type of thing where as you did these projects, each individual projects had its own nuances but you had to frame things in a way that again at the time you didn't. I didn't know we were doing this, but they were from a B side perspective. So, for example, going back to some of our own old sayings, unseen is unsold. So that's kind of talking about the whole idea of breakthrough on shelf. And again, a behavior that a lot of times consumers would tell you one thing, but then you'd observe them and they're doing something completely different. And you have to tie those two pieces together. So when you start thinking about system one and System two, it's basically that looking at system one behavior, basically understanding how consumers are acting or reacting to certain stimuli in store, and then what they tell you that they actually would do or how they feel or what, whatever it may. The case may be, could be different, could line up. But you really. It's about tying those two pieces together and telling one cohesive story that is going to essentially help our clients understand what they need to do and make those good business decisions to push their businesses forward.

Right.

Well, and like, as you said, it's not. We really couldn't have grown up with this dream of being a behavioral scientist because it just wasn't a publicly known field. It didn't exist.

Exactly.

I've talked about it a lot on the show that. So I got my undergrad in marketing and business administration and worked in branding. And there was like this tiny little bit in one class on buyer psychology. And I was like, yes, like, I want to do this. This is amazing. And I spent 10 years calling universities that said not a thing that doesn't exist. Sorry. And just not knowing what the path might be and what was brewing up in the psychology departments or an econ space or just so many fascinating. I love hearing how people's paths have made here, you know, where like engineers that ended up pursuing behavioral economics. So just is a lot of fun to kind of hear in a world where it's being created as we're. We're all making it. Right. So getting to see all those paths and how it comes together, and I think that's really helped to shape the field in a way that there's so much insight from all these differing spaces. The brain is so complex.

I just.

I enjoy that. No, there's not just a set path that you have to go through to be a behavioral scientist. I think that's great.

No, absolutely. It's. It's something that can be applied to pretty much every aspect of our lives. And it's happening unbeknownst to us every day and every decision that we make. So it's just taking that and just getting A better understanding of how those decisions are made. And so I think it's just generally interesting for most people and we just happen to be applying it to market research.

Yeah. So let's then talk a little bit about. So this future of people are talking about like machine learning and AI and what that's going to be in behavioral science. And knowing that's. I know that it was on a docket for 2020, many conferences that were set up originally to be really focusing on that. Of course there was a little hiccup that happened in that process.

Just a little bit.

Yeah, yeah. Just a little shift in all of our plans. But you all at behaviorly have Flash AI, some really cool stuff you're doing there. And because I know this AI side is so top of mind, pun intended, I suppose tell a little bit about what flash AI is and what cool research y' all are doing.

Yeah, Flash AI is basically. So as you mentioned, the industry has been moving on this trajectory towards machine learning and more broadly leveraging AI. I don't know that there's been anyone yet who's kind of cracked the code. And because it's so very early in its inception, it's kind of the wild west out there. So it's just figuring out how to put something together that's actually not just a marketing gimmick that's really going to pull in real BSI into it. And also on the back end is going to be predictive. Right. Because in the end, that's what our clients need, is something that's going to give them the right answers to the business questions that they have in the long run. So what we did was we came up with a system called Flash AI. And the key to our system was this couple of things. First and foremost, going back to this history that we have, we have a database. And that database is not just a bunch of numbers sitting in a repository. That database is actually broken up into system one behaviors and system two behaviors. We're able to take those learnings and apply them to the machine so that the machine can basically learn how to mimic those behaviors. That's the first piece that's really critical to at least our system working. It's based on previous behaviors. Number two, our system applies image recognition. Right. So I think it's like 80%, if I'm not mistaken, of what you take in in the world is through your visual. I mean, you do have other senses, touch, smell, so on and so forth. But most of it is you're reacting to visuals. So our System basically takes that approach and the machine basically reacts to the visuals in a similar way as what a human would do. And so we take those two pieces, put them together so that the machine is basically reacting to those visuals via image recognition and then the database, because we know what happens when you see certain types of visuals or elements or stimuli. And we're able to then predict what that potential, whether it's from a breakthrough perspective, what's likely to happen from a shopping findability perspective, or even if it's from a communication perspective, we're able to essentially predict what's likely to happen leveraging the machine learning. So in a nutshell, it pulls all of those pieces together and allows us to really have a cohesive platform that takes into account all of the key pieces that really, from a behavioral standpoint, drive what consumers actually do in marketing. And just going back to your earlier point about just a little bit of a hiccup and what was happening over the past year, actually that's something that just the world has. It's changed, but the change that was happening was already kind of on a changed trajectory. And all Covid did was essentially accelerate that change. And so as we're thinking about things like other applications, e comm and so on and so forth, all of these things are just going to be more important. In the new world where we're seeing different types of shopping behaviors, it's not just about going into the store and seeing the pack on shelf. So this is just, at least for me, this is kind of foundational work to the future of where we're going to need to be. And again, we've seen how it's changed in just the year. Who knows what's going to happen over the next year or two?

Right?

Well, I know. So I had the honor of being the emcee at an event that was put on by behaviorally and a couple of other great partners called you'd better behave. And knowing that was March of 2021, but having the, you know, researchers talking about, I know you talked about framing earlier, but this reframing things as the opportunity and in the world of always feel like you still need the caveat of, of course we wish it didn't happen and we weren't here, but as we are, you know, you had that opportunity to say, you know, in this force shift where status quo is gone and people are changing behavior, what do they even do? Like, did they go way outside of a pattern? How does that become predictable when we're now forced into having to Buy groceries and having them delivered. And what does that do when you're not going through the checkout anymore? You know, and all these research questions that you're able to look into if you're taking the time to look for them instead of just going, oh, no, I guess everything's on hold. Yeah.

The world is just becoming increasingly more. I mean, it was already digital, but it's just becoming increasingly more digital. And that's why we focused on. Part of our focus is on tools that are going to allow us to understand the landscape now and into the future. So we're just trying to position ourselves in a good place, whether it be from kind of your traditional research standpoint and how we look at things and how we tie things together, but also with the tools that we bring to market.

Yeah, well, and I just remember the different researchers that were presenting at that event, which was about six hours long, of which I was on, you know, for the whole six hours. Yep. But with that, you know, they're talking about, you know, this is the second Easter where we're going to be, you know, and it's going to be the second, you know, Halloween, potentially, and 4th of July and just all these things where, you know, you're going into a next summer and what can be learned and applied. And so I am curious, I want to talk about some of the case studies you have from Flashai. There are some really cool ones within there. And perhaps it's a case study and then answering this question, but I don't want to forget it. With the research and knowing that the machine is doing learning, of course, and it's programmed to be like the human brain, but as exposures change, as we get used to packaging or, you know, things along those lines, are you having to kind of in tandem have actual people doing the eye tracking and research and then that is continually feeding the machine to just make sure it's updating as you go, or are we relying on what the machine thinks will happen?

No, it's actually. It's more the former. So it's a dynamic system where the existing work that we're doing is feeding the future work and the future models that are being built. So the models are continuously being updated to essentially reflect the real world. So that was a key piece to us. We don't want our number one database to be static. We don't want the models to be static because we know the world is changing and it's changing pretty rapidly. So absolutely, we make sure that those models are continuously being updated and Reflect the real world as it is today.

Yeah, so that's the. You have the machine that's going to test and say, we're pretty sure this is what would happen if you go and do a launch. And then when you launch, potentially test it with real people to see what actually happened and then can kind of feed itself in that process.

So that's more of the kind of like a validation is what you're describing. So there's two pieces to it. There's the validation work to make sure what were the models are actually accurate and that they're reflecting, you know, what would actually happen in market. There's also taking information that we've tested and that we are testing and feeding that model so that essentially if there are certain categories are being shopped differently now because of what's happened over the past year, for example. And so if any of those behaviors are changing, the system will essentially be able to pick up on that and we'll be able to tweak those models to make them accurate. So it's twofold. It's essentially validating the work that we've done to make sure it's accurate. And then there's just this dynamic aspect of making sure that the system is continually keeping up to speed with or at the pace of change that's happening in the world.

Knowing that you had the line of years of this unseen is unsold, and going through a couple of the case studies where that product packaging and standing out on the shelf, it looks like, is a lot of where your focus is. So can you give a couple of examples of some of the case studies, whether it's one of your favorites or most remarkable finding or whatever that is. Let's hear about them.

Well, I mean, I guess the key thing is a lot of what we've done is basically looking at. It's not just the packages that we're testing, but also the clutter or the context around those packages. So that really drives, particularly from a breakthrough standpoint, that really drives how well something works. I guess from a remarkable standpoint, the thing that I'm most proud of with this system is it's not just looking at the test design, it's looking at the context around it, because content, as we all know, context matters. So you can have a bunch of. I'm just making up purple packages surrounded by purple packages, and it may not break through as well, whereas. And certain. Certain varieties within the line may not break through as well, whereas you. If you have pink package in a sea of blue it's going to actually break through pretty well. And so the way we built the system is not just to focus in on the test packages, but that context around it and take that into account as well. Because if you don't take the context around it into account, you put yourself in a position where you can either get false positives or false negatives and essentially give the wrong direction to our clients. So that's probably the thing that I'm most proud of with the system is just the ability to take context into account. The other piece of it is looking at the individual packs that we're looking at. It's not just about, I guess, basic elements. It's also about tying those elements back to other similar elements from things that we've tested in the past. So, for example, branding. For example, you can have huge brand mark and, you know, for example, a candy package where you have a bunch of different. A bunch of individual candies versus having just one piece of candy or a character. The system takes that into account and is able to differentiate and then basically tie that back to, well, this. This worked in this context, or it didn't work in that context. And it allows us to basically get very pretty fine detail as far as how well something is likely to work or not work. I think it's just the nuance of the system is probably the thing that I'm most proud of, because again, context and nuance makes a difference when you're talking about not just how the system evaluates things, but behavior in general.

Yeah, absolutely. And I have made some notes here that will link in the show notes to the episodes on relativity and anchoring, because that's a little bit of, you know, in our mind, we expect that we're going to be seeing purple packages. And so then when the pink one is there, or, you know, blue to pink, like you were saying, it's going to stand out in a different way because of those factors. I also am linking to the episode on color Theory, which talks specifically about red buttons on the Target website are not going to stand out in the way that they would on a Facebook website because of what we're seeing everywhere in those dominant colors. And so that's so important. You also talk about in the case studies, you know, so there was some insights. There was a kombucha that had short bottles. And so it's saying, well, if everything else is tall, you may think, well, we want to be short because we're going to stand out. But maybe it doesn't. We work quite that Way also knowing that the different, you know, shape of the packaging on, I think it was a deodorant where it's, and you know, it was only going to be at Target. And so like this was a good real life test because, you know, these are the shelves it's going to be on. This is what things look like at Target. We can match that versus if you're going to be at, you know, Target and Walmart and Fred Meyer and everywhere. It's, you know, there's a lot of variables that come into play there. So if you want to walk us through a case study, client comes to you, what's the type of question they're looking to answer? How do you help shape that and turn it into a research study? Walking through that process, you know, what you're able to find.

So generally speaking, we see the Flash AI tool as an early, earlier stage tool. So generally those questions that are coming, they're coming to us with, are more upstream. So it's more about, we have these designs and we're considering either making a change or we have this new product. So the Flash AI helps a ton and from that perspective, so that you can essentially test without having to spend the time, money, energy into talking to real consumers, but getting the same type of feedback. So those types of questions are the ones that clients come to us with sometimes like in the case of the raw sugar, they involve structural changes and as we mentioned, just from a, from an overall look and feel standpoint, those, those designs can sometimes be close in further out. It really just depends on what their challenge is. So basically it, that's why it's critically important for our design, particularly when it's talking about new products, new unique products, to understand that breakthrough at shelf, how those packages are going to work when they're essentially standing in a sea of clutter versus the consumer has it in their hand and they're picking it up and they're making that last decision essentially whether to put it in the, whether it's the virtual cart or whether it's the physical cart to take it home and to, or to purchase that item. So that's generally how it works. And when consumers come, when clients are coming to us to get learning on their packages using Flash AI.

Yeah. And so it seems, is it accurate to say that you're able to test this before they get too far downstream and maybe have to create a bunch of packaging and you launch it and say, well, something's wrong, can you help? What did we mess up? And then you have already pre printed umpteen zillion packages that you have to get through or you throw away and have the loss on where because you're using the AI and other virtual techniques, you're able to test a mock up essentially of what they're thinking about putting out. And to say, you know, to use the kombucha to where we say, okay, yes, the being different is important, but short is not where you want to differentiate because it's not standing out, or there was the, you know, the act mouthwash and where it's great to be having something without a bunch of dye in it. However, you don't see it because it's clear. So something needs to happen so that it can stand out against its competition.

That's definitely the approach. It's not just about when they get to the final, I guess to the end and they're just, they're launching and keeping them from launching or potentially saving them from launching something that doesn't work well. But it's also about getting to the optimal design that's going to work. That's not only going to work in market, but also a lot of times what we see is we test, then we end up retesting and there's a lot of work. And a lot of our clients time is not just money, but it's even more important than money because you're dealing with short turnarounds, you're dealing with needing to get in market pretty quickly and tight timeline. So basically the flash AI lets you essentially evaluate those upfront, not have to do a bunch of rework. And then when you go to market, be confident that you know what you've evaluated is really going to have the impact that you, you intended. Also what goes into the research, because we're not saying the other thing just to be clear, we're not saying this is to replace the final validation. This is really the first step. And then hopefully it siphons down the different options that would go into a full test and then you can fully validate those and essentially do any fine tuning that you need. So it's, it's just improving your odds of success in the end. I mean, I've used the analogy of the gambling analogy to basically make this point as well. Because a lot of times our clients are to some degree gambling just in the back room figuring out, oh well, this person likes this design, my gut says this design is better, but to some degree that's just gambling without having a sense of what may or may not work. The idea is to up those odds of having something that will work by having data to go based that on. And then if you think about the gambling analogy, then you start to play with house money basically when you're doing that. And it gives you just a much better chance of success in the end. So that's the thinking.

Yeah, I like that analogy. The one I came up with when I was writing my book was this. So it's like. So that going with your gut is throwing noodles at the wall to see what sticks. Right. And so, you know, I don't know where at one point, but it's like properly applied behavioral economics allows you to throw darts while your competition's still throwing noodles. Right. You got. At least it's gonna, you know, it's gonna hit the board somewhere. Right. You're getting closer. You can work on your aim as you go.

Exactly, yeah. It's the same idea.

Same idea, yeah, absolutely. So as you look at the future, like you said, it is sort of the wild west in AI and this machine learning and knowing what that means for being able to more accurately predict behavior. If you think about the next five years, what gets you excited? What are you all looking to be able to incorporate or just what do you think people should be having? Top of mind in this area.

Yeah, I mean AI isn't going away. It's just a matter of what that looks like in the future. I mean, I think there's still going to be a place for real consumers to give feedback because in the end that's what we're trying to replicate. It's just where that's headed. So I mean, E. Com obviously is a big part of that. And as that grows, I think the natural shift is going to be more towards that. When you start looking at places like what's happening in China, where I believe this year, basically they're the first country where over 50% of their retail sales are going to be online this year. So with that kind of shift, you're going to see a paradigm shift in how we need to measure consumer behavior. And it's just, I think that's going to be one piece of it. Virtual tools, I think kind of connected to AI are also going to be another piece of it. Particularly as people shop more online and use either are shopping in non physical environments more and the other piece of this is also non physical environments or tools are impacting consumer behavior as well. So that path to purchase is changing as well. It's not just about where they're shopping and kind of what they're buying, it's about everything in between. So that path to purchase is going to play a critical role as well in what we see in the future. But I think technology, you can try to run away from it.

Good luck.

It's something you have to. Yeah, it's something you just have to build into what you're doing and leverage it. It's using, I mean, the old kind of old tools, the tried and true tools are, you know, they're great, they play their role, but it's just as we move into the future, it's. We're just, we just need to tweak some of those to make them a bit more relevant for the world of today. And I think AI, I think virtual, maybe not virtual. I think in the past virtual had been this very heavy, let's replicate a real store and all of that. When I say virtual, it's more about just a digital environment and digital shopping, what's happening online. So that would be my guess. And again, I think AI is going to play a pretty big role in that. We're already seeing it.

Yeah, well, and that's. I know pre pandemic, I was working with clients on the, you know, as things are changing of what's. Be able to be at the checkout with regulatory stuff and you know, if you lose the impulse buy at checkout for some brands, I mean, that's, I won't say everything, but that's a huge place of, you know, you're not being purchased anymore and then what do you do? So that was already being discussed well before pandemic times. And then now to say, you know, it's not like everybody's going to start. I mean, I guess maybe, but probably very unlikely that everybody has a virtual reality headset. It's like time to go to the store.

Boop.

Like going to pretend like you're walking around, which like, I'm sure you know, you know, and have used these where you're doing the VR testing to see, you know, I'm going to grab this off the shelf and be able to experience it because you're trying to emulate that experience. But when the context is totally different, when I'm shopping through an app from my preferred vendor, what's that experience going to be? How can you be incorporating those, the context and with. Without end caps, you know what happens and these associations of bringing things together. There's a lot that is lost when you switch to those digital environments. However, as we were saying with the reframe, you know, there's always Looking at it as an opportunity, what could we do? If we take a step back, how can we make this an amazing experience for everybody in a really powerful way?

So, yeah, I mean, when you, when you going back to China, I mean, there's looking to the future beyond China, but because China is so much ahead of kind of the curve on some of this stuff, I think that that's kind of the near future for the US Market and probably parts of Europe as well. Things like M Commerce doesn't get right now, we don't have the infrastructure that perhaps places like China have. But as that infrastructure is built up, those things are going to change pretty rapidly. So it's just, it's the clock is ticking, basically.

Yeah, absolutely. Well, and when you have that virtual space, you know, both the purchasers and the vendors can really be from anywhere. So yeah, like you said, clock is ticking. Your competitors aren't just those next door anymore. So lots to do there. Well, for everyone who wants to learn more and, you know, connect with you and learn about Flash AI and behaviorally.

You know, what's the best, what's the.

Next step that they should take?

Well, they can reach out to me directly@sam.alberthaviorly.com that's one place our website is a good place to start if you're just perusing or just want to get a little bit more info and just reaching out to us directly. It's pretty simple as that. We're here to help you and happy to wherever we can.

Awesome. Well, I will of course link to the behaviorally website. I'll link to some of these case studies that we Talked to using FlashAI as a reminder to listeners to save our guests email inboxes. I will not be actually typing out and leaving a link to Sam's email address on the Internet because nobody needs all that spam. So that's one to take a note. It won't be in the show notes, but it exists and you can be ahead of the curve to go ahead and grab that. So once again, Sam Albert, thanks so much for joining me. It has been a lot of fun and interesting to hear, you know, what y' all are doing on that forefront of AI and behavioral science.

Yeah, absolutely. Again, thanks for having us and look out for more to come on the AI front. There's stuff cooking in the kitchen right now.

Ooh, exciting. I look forward to hearing about it.

So what got your brain buzzing as you learned about how artificial intelligence and behavioral economics can work together today? For me, I really love the intersection of technology and and behavioral economics, how we can incorporate artificial intelligence to do better research and help brands to leverage insights faster and to do that in ways that saves time and money by helping to discover what people will be more likely to buy and what isn't going to convert. This type of research can reduce the need to print a bunch of boxes or displays or containers to run physical tests, which is good for sustainability purposes. And remember, this conversation originally aired back in 2021, so the technology is only continuing to get better. This can be for things like virtual tests and for digging in on insights along vast amounts of data, which you can hear more about in my conversation this upcoming Friday with Hunter Thurman, president of Alpha Diver. He's going to share about some fascinating insights they've uncovered that have to do with pricing and how, as I always.

Say, it's not about the price.

You don't want to miss that one. So if you aren't already subscribed to the Brainy Business podcast, now is a great time to do so. As we close out the show, don't forget about those show notes with links to my top related past episodes and books and more. It's all waiting for you in the app you're listening to and atthe brainy business.com 4. And just like that, episode 440 with Behaviorally's Sam Albert is done. Join me Friday for a brand new episode with Hunter Thurman, President of Alpha Diver. It's going to be a lot of fun. You don't want to miss it. Until then, thanks again for listening and learning with me, and remember to be thoughtful.

Thank you for listening to the Brainy Business podcast. Molina offers virtual strategy sessions, workshops and other services to help businesses be more brain friendly. For more free resources, visit thebrainybusiness.com.