Why Elite Supps Chose Behavioural Science Before Bargains with Behamics | #608

Brooke Eichhorn from Behamics joins Johan Nyberg, Chief Digital Officer at Elite Supps, to unpack how one of Australia’s fastest-growing supplement retailers is increasing conversion without leaning on discounts. With over 140 stores and rapid global expansion, Elite Supps isn’t short on traffic: the real challenge is helping customers make better decisions once they land on site.

In today’s exclusive episode, the conversation goes deeper than traditional CRO. It’s not about tweaking buttons or shaving seconds off checkout. It’s about understanding why customers hesitate, and using behavioural science, powered by AI, to respond in real time.

Today, we're discussing: 

• How Elite Supps increased conversion without relying on discounts
• Why most ecommerce nudges lose effectiveness over time
• The difference between static CRO tactics and dynamic personalisation
• What “causal AI” means (and why it matters for ecommerce)
• How to reduce hesitation instead of just chasing more traffic
• Why first-time vs returning customers is too simplistic
• Where the real revenue opportunities are hiding in your funnel

Explore Behamics
Connect with Brooke
Explore Elite Supps
Connect with Johan

Transcript

Introduction

SPEAKER_00 0:01

And if you're not using them, like you are leaving revenue on the table.

SPEAKER_02 0:05

We can increase conversion rate and average order without pulling that discount lever. That is gold for us.

SPEAKER_00 0:12

That one over there was having a negative impact. It was worse than doing nothing at all. Hi, I'm Brooke Icorn from Behemoths. And today we've got Johan from Elite Socks on the line with us, and we're talking about how we've been applying behavioral science dynamically with AI to grow their site conversion.

SPEAKER_03 0:32

Hello, hello. My name is Bushy, and you are joining me on a very special edition of Ad Descartes. Joining you from the land of the terrible people here in Brisbane, Australia, as normal. But today we have a bonus Wednesday episode brought to you by our partners at Bahamix. Now, most of us in e-commerce spend a huge amount of time and energy optimizing by testing headlines, tweaking checkout flows, adjusting button colors. But how much of that optimization is built on a real understanding of how the human brain actually makes decisions? Or is it guesswork? Because every customer who lands on your site is carrying a different level of certainty, a different level of distraction, and a different emotional state. So why are we building one experience for all of them or the same experience for them every time? Our two guests today are showing how to develop a dynamic experience, taking all of that in mind. So I'm lucky enough to be joined by Brooke Icon, who is the APAC country manager at Bamix. Bamix is an e-commerce optimization platform that has spun out of the University of St. Gallen in Switzerland, founded by a behavioral scientist and a former Google Fellow. Now, before joining Bahamix, Brooke spent nearly 13 years at eBay across senior roles, including head of retail marketing and head of fashion, luxury, and authentication. So she's seen consumer behavior at some very serious scale from both sides of the marketplace. Joining Brooke is Johan Nijberg, who is the chief digital and e-commerce officer at Elite Subs. Elite Subs, if you don't know them, are Australia's fastest growing supplement retailer, now with over 140 stores nationally and expanding into Singapore, Dubai, and New Zealand. Johan even owns his own Elite Subs franchises, which gives him a genuine full circle view of how digital and physical retail reinforce each other. In this conversation, you'll hear how Bahamic's AI model trains specifically on your customers in real time, not generic internet data, and why that distinction matters. You'll hear how Elite Subs drove meaningful conversion lift without discounting. We all like to hear that. Why first-time buyers need to be treated completely different to returning customers, and how Johan built a tech stack where every tool has a very distinct job and they all feed each other rather than fight for attention. A massive thanks again to Bahamix for their support during 2026. Now, here's Brooke Icorn and Johan Nijberg to take us away on this very special behavioral science episode of Ad to Cart. Brooke and Johan, welcome to Ad To Cart. Hey Nathan. Brooke, I'm going to start with you because we had the privilege of having Sonia Friedrich in with us. It was almost two or three years ago, talking about nudges and behavioral science. And we got talking about Bahamics. And that's when it piqued my interest around how do we use customer psychology in our front end to really increase those conversion rates. But it's come a long way from then. I want to know from you, how did you get into Bahamix? Because you've got such a fascinating career yourself.

SPEAKER_00 3:59

Damn, interesting question. Yeah, how did I get into Bamix? My background is in e-commerce and marketplaces. Look, a big portion of that was at eBay for, you know, nearly 13 years. And over my time there, it was very much two sides of the coin. It was running retail marketing and retail promotions. So going through the journey of having big retailers like The Good Guys, Maya, Country Road, right through the e-compure plays like Kogan and Catch and Scaling Then on site. And then flip the other side of the coin, had been running circular fashion and C2C for quite a while as well. So yeah, a lot of behavior change, but also how do we drive change outside of discounts and promotes?

Elite Supps x Behamics

SPEAKER_03 4:43

I love that. I think everyone listening will be going, tell me the secrets. How do I do it without those discounts and promotions? Speaking of that, Johan, Elite Sups, I was in one of your stores on the weekend and it was a really great experience. Top-notch customer service. In at the North Lake store, I was escaping for a little while while my wife was in at IKEA. So I was like, I'm gonna go check out and hang in Elite Subs for a while. Tell me about your journey. How did you end up at Elite Subs and what's your role there?

SPEAKER_02 5:07

Yeah, so currently I'm the CDO of Elite Subs, but my background is digital marketing and e-commerce. I used to run two digital agencies back in Europe and then an e-commerce business back in Europe. Then uh I met my Australian wife in a fitness camp in Thailand and got tricked to uh move to Australia and here I am.

SPEAKER_03 5:28

See, we're most people make bad decisions in Thailand, but yours was at a fitness camp. So I'm assuming yours is a little bit more, you know, organic and and natural.

SPEAKER_02 5:37

Yeah, and uh definitely not something I regret. So uh no, it's been uh it's been great. But then you know, taking the decision to move to Australia that had impacted on my European business. So basically I stole them and then uh started looking for a job. And you know, e-commerce is not huge here in Canberra, so um I thought, okay, that's gonna be tough. But then Elite came on the radar and uh you know it fitted perfectly with my own values and what I'm interested in. So yeah.

SPEAKER_03 6:05

Amazing. And we were talking just before we jumped on the call that you're not just CDO there, you actually own your own franchises as well.

SPEAKER_02 6:12

Yeah, I've been uh lucky enough to uh have a few. I got presented an opportunity, and it's just a fantastic way of getting that full 360 of the business because you you know running e-commerce and digital is one thing, but then seeing the operations from a retail side of you is something completely different.

Customer Experience and Service Excellence

SPEAKER_00 6:30

Well, they're popping up everywhere. Nate, I saw your selfie on the weekend at Elite Sops, and you literally sent it at the same time as I'd walked past one of the franchises that is just opened in my local store. So yeah, they're popping up everywhere, Johan.

SPEAKER_02 6:44

Yeah, we're about uh an incredible growth. I think was it 2025? I think we opened like four to five, fifty stores just in that year. It was like a new store every week, and we're now at just over 140 plus here in Australia. We have two in Singapore at the moment, soon to be four, opening up in Dubai and New Zealand very shortly too. So it's a lot happening.

SPEAKER_03 7:04

What makes Elite Subs special? What is it behind the growth?

SPEAKER_02 7:09

I think we have a purpose. We want to really help our customers in their uh fitness journey, and we take great pride of it. Everyone who works at Elite are certified nutritionists, and we talk to customers every day about their goals and preferences, recommendations, whatnot. So I think the level of service from our whole team and the way we care, I think that's a big piece of it.

SPEAKER_03 7:30

That's a big commitment to get certified nutritionists in.

SPEAKER_02 7:33

Oh, it took a lot of effort to uh, I think close to 600 or 500, uh well over 500 people at least. So when you reach our customer support, for example, all of them are certified. And then um, yeah, as I said, our store members are uh too. So amazing.

SPEAKER_00 7:50

I never knew that about that your business, the certified, yeah, that's incredible.

SPEAKER_03 7:55

Yeah, it's been a bit over a year now. That's cool. And with such a big expansion in the physical footprint, what role does e-com play for you, Johan?

Understanding Nudges and Behavioural Science

SPEAKER_02 8:05

Retail is definitely uh the biggest side of the business, but e-comm definitely you know fits it part there now with uh click and collect. And we have some other areas of new products coming up that will tie in nicely to the Omni-channel experience. But e-comm also, I think a lot of customers go on our site to explore and then before they go into uh one of our stores, so it definitely plays an important role there too.

SPEAKER_03 8:27

Yeah, and Brooke, we just heard from Johan around how much they value customer experience and actually having real knowledge on sites when people contact them. How have you seen it from a behavioral science perspective? What sets elite subs apart from everyone else?

SPEAKER_00 8:42

Yeah, well, from a behavioral science perspective, yeah, Johan came with that brief of our customers have goals and they can be very different goals. Some customers might want to be losing weight, you know, shredding, and some might want to do the opposite, bulking up and putting on weight. So, you know, very goals-based, but also wanting to bring the digital experience forwards and be more dynamic. So, you know, they'd been experimenting with behavioral science before we met. Things like, you know, hard coding how many items had sold, you know, in the past month on things. But Johan wanted to take it just another step forward than that and get a bit more intelligent with what we were doing behind the scenes.

SPEAKER_03 9:22

All right, we're gonna dive all into that and I can't wait. There's so many parts to that statement that you just had there, Brooke. But where I want to start is around nudges because when Sonia was with us, my very simple brain around behavioral science could understand nudges, and I think it's a great gateway in for anyone looking to know how this actually works on site. So, Brooke, like a lot of people think that they might be using nudges, whether that be countdown timers or hey, there's 17 people looking at this item right now, or pop-ups that happen just when you're about to exit. They're all like little nudges of behavior. Is that what you consider great examples of nudges?

SPEAKER_00 9:59

Yeah, well, those examples there, they're examples of social proof and and scarcity. And they're really well used in e-commerce, you know, heavily used in in most sites. And it's because they are effective, you know, really effective at driving behavior and behavior change. You take something like social proof. There's a lot sitting under the hood of that. You know, we look to others to see how we ought to behave. And whether we're looking for social norms or herd effect or bandwagon effect, you know, it can really be that subtle subconscious influence of behavior. But where the effectiveness starts to erode is if it's overused. So if you say you've got a site and every item has there's this many items left, or you know, this many people have bought it today. Customers just start to distrust that. It looks kind of hard-coded and not genuine. Or maybe if you're going to like a hotel booking site and I don't know, every listing has two rooms left.

SPEAKER_03 10:53

Yeah, exactly.

SPEAKER_00 10:54

You know, the consumers, the sniff test on that, starts to so that's where the effectiveness really starts to peter out. So we find that a dynamic approach is always best. So not just dynamically pulling in numbers, intelligently targeting the nudges themselves, because we need to use the right nudge for the right customer at the right part of their visit and even down to the right skew level. So, yeah, that's where we aim to get. And there's a whole wide world of behavioral science out there that you can tap into outside of social proof or countdown timers. And if you're not using them, like you are leaving revenue on the table.

SPEAKER_03 11:27

So not one nudge for everyone and don't overnudge.

SPEAKER_00 11:30

No, not one nudge for all.

SPEAKER_03 11:32

How does that align for you, Johan? How do you make decisions on when you give those little kind of hey, upsell here or you know, urgency here? How do you strike that balance?

SPEAKER_02 11:41

Just as Brooke said, you you probably because before I met Brooke, we had hard-coded like, you know, this product has sold this many uh items per month. And we did it exactly in the the way that Brooke said, it was on like on every uh product. And I think it gets a bit messy. And that's what I also started to feel okay, this is a little bit messy. So I I think Behemics approach where it's dynamic and intelligent, that's the way to go. And then how you do it? Well, you need someone like Behemics to be able to do it. And one of the main reasons, I guess, why I haven't added a um Shopify app where you have those popping up in the uh bottom left corner that you know Josh W from you know South Australia bought this 15 seconds ago, and it doesn't matter if you, you know, come in incognito or if you uh you know uh have the same session, it's it's always the same message and it, you know, uh people see through it.

The Science Behind Behamics

SPEAKER_03 12:35

It loses trust, doesn't it?

SPEAKER_02 12:36

It definitely does.

SPEAKER_03 12:37

I suppose that's really hard too with you because you've got such a dedicated team of experts and customer service there by just throwing up random nudges, and I'm putting nudges in like quotation marks there, can erode that trust and that belief that there's real people behind this.

SPEAKER_02 12:52

Yeah, yeah, definitely.

SPEAKER_03 12:53

Brooke, tell me more because we're going to get into the science of it, but I think it's probably important that everyone understands where Bamix came from. Because it was a fascinating story built by behavioral science and a Google fellow. Tell me around the philosophy of Bahamix. How do you ground everything in science-based versus just science-inspired? Because I could imagine it's it could easily be a form of greenwashing. Let's call it science washing. How do you make sure there's real science behind it?

SPEAKER_00 13:24

Well, we're literally a spin-off of a university in Switzerland. And I co-founded TLO. He did his PhD in behavioral science for e-commerce. So that's where he specialized. And he's still lecturing. Love working with this team. So yeah, he's still lecturing in in uni's on just that subject matter. So yeah, there's a robust depth of you know evidence-based psychology that's sitting behind it, which means that, you know, whether we're working with supplements or, you know, trying to sell cars that you know we can apply in a way that is robust. And then I guess on the other side of things, the AI, you know, when Tilo met Valin, our other co-founder, yeah, Valin had exited an AI company already, was a Google Fellow, and they've been trying to solve a similar challenge. So we're developing the AI itself using scientific methods. So we're training the AI models in causality. So, look, in the world of science, say, like a clinical trial or an economic trial, you need to back out the causation of the interventions, and that's exactly how we're training our models. And then, you know, because they are in the science and academic world robustly testing everything we do, we need to prove every dollar of uplift we deliver for a client. So taking the world of randomized control trials into e-commerce, you know, has just been a way that seemed completely normal to our founders, but I think is quite unique coming from an e-commerce background.

SPEAKER_03 14:47

It's very unique. Johan, did you understand the science when you signed on?

Implementing Behavioural Science in Ecom

SPEAKER_02 14:51

Not really. But anyway, the very first time I came across Behemoths was in April last year at iMedia. And I just saw on the schedule that uh, you know, a company called Behemoths was going to uh have a session, and that name just stood out. I love the brand name. It's interesting and you know, it's curious. So I started to read through it a little bit more and the way they do like intelligent artists that really stood out, and then but the all the science behind it, that's just fascinating. And you said you sounded dumb. You know, imagine you're sitting in the room with the co-founder and then uh a couple of data scientists, and I've been like, okay, I am the stupid guy in this conversation. Like I'm here to sell more protein powder, guys.

SPEAKER_03 15:35

Yeah, for your workouts. So, where do you start then? I could imagine that you guys came together and was like, hey, there's a huge opportunity here to be smarter around how we communicate with our customers and be more personalized. How do we even break that opportunity down? Where do you normally start? Brooke, I might ask you first to go, where do you normally kick off in Bahamix?

SPEAKER_00 15:54

Yeah, we usually kick off by understanding the client, their customers, their goals and strategy. Because yeah, we're not like an off-the-shelf tool where you go in and pull some widgets in. We literally have a team of behavioral scientists that come up with the initiatives. So, you know, they're using evidence-based science, what we've learned over years with with other clients. And then it really is a collaborative process. But the real learning kicks in is like when we deployed a site and the model starts learning. So the model is self-learning from Johan's customers. We don't want to train the AI model on the entire internet or you know, another site. It's then don't come with preconceived notions. No, so yeah, we really need to understand what's resonating with the customers where things are backfiring, and we need to be able to adapt really fast because it's all about how do we just drive revenue when the customer's on the site.

SPEAKER_03 16:47

And how long was it before you started seeing that model learning for you, Johan?

SPEAKER_02 16:52

We deployed uh Behamex, I think, beginning of July, Brooke. Was that right? Yeah. Yeah. And then I guess we had about the 30, 40 days where you could actually see the model was, you know, learning the audience, testing heaps of different nudges from there. So I guess mid-August, you could really start to see the uplift.

SPEAKER_03 17:11

Good timing before Black Friday.

Measuring Success and Uplift in Sales

SPEAKER_02 17:13

Yeah. Coming into September, which is a you know historically a pretty good month for us. But then we also have what we call our elite week, which is a very big sale event for us. And as I said, we started to seeing uplift in August, but then in September it it kind of really took off. And for me, especially during the sale event, when I logged in and looked at the Behemoths dashboard, it was like, all right, oh geez, what's happening here? It was so it was a bit of a flashback to you know, you log into Facebook or TikTok uh dashboard and look at their attribution models, and I'm like, all right, something is a little bit too good to be true here. And I I try to wrap my head around, you know, what's going on and what am I missing? So I actually paused behaving after September because I needed to validate the numbers. It was yeah, it was that um big of an uplift in in uh September. So thankfully, we had just completed a huge product uh to integrate the Coveo. And the company who did that integration with us, they had a couple of data scientists. So I asked them to take a look at all the Behemoths data that Behemics provided and our Shopify NGA4 data just to validate and you know, is it statistically significant or is it just noise in the control group or what's what's happening? And uh yeah, it came out it's true, it's accurate.

SPEAKER_03 18:38

It was legit. It wasn't a Zuckerberg figure, it was a legit figure.

SPEAKER_02 18:41

100%. And you know, that was very refreshing to actually see that because uh yeah, the numbers were uh incredible. Um Brooke you can go into more detail around what happened during that specific sale event because that's also interesting. What type of nudges that actually deliver results on a BAU day versus a sale day?

SPEAKER_03 19:01

I'd love to know what those nudges are, Brooke. And and also what are the key figures that you're looking at when you're looking at that dashboard? Like, is it conversions, is it revenue, is it GP? But Brooke, maybe if you can give us a little bit of an overview first on on what that initial, you know, those unbelievable stats are that you scared Johan with. Well, what are we talking there? What kind of executions?

SPEAKER_00 19:19

Yes, when when Johan said that, we went, okay, we're here with you.

SPEAKER_03 19:23

We've never lost business because we were too good.

SPEAKER_00 19:26

Yeah, like it was it was quite a unique case. And yeah, he had wanted to do this this third party review, and we were like, yep, here's everything you need, here's the data. And I think we were quite confident in our methodology. But yeah, what numbers do we look at? It's ultimately revenue and conversion. We've got lots of other metrics that sit under that, but the AI model itself is geared towards driving up conversion and revenue per visit. So they're the main metrics we look at. And look, the whole reason we have this dynamic approach is so that we can respond to the customers as they're behaving in that week on that day. So we did definitely see different behaviors showing up in elite week. Things like, I guess customers were really responding to the specificity of nudges. So things like, you know, this ex-visitors bought this in the last 24 hours, that uplifted the revenue per visit by like over 50%. So that one performed really strong. But then we see nudges that are quite similar, backfiring. So it was like good choice added 10 times today to cart was having a negative effect.

SPEAKER_03 20:29

What do you mean? So when you add it to cart and you say good choice, yeah.

Adapting Nudges for Customer Intent

SPEAKER_00 20:33

So because we're running multiple nudge initiatives, we're not picking a winner and hard coding it, just simply the difference between, you know, this many people have bought it in the past 24 hours versus showing good choice. It's been added, you know, 10 times to cart today. That one over there was having a negative impact. It was worse than doing nothing at all. We need to understand with causality, you know, what are the backfire effects, what are the counterfactuals. What if we did nothing? Would it be better? So, you know, and Elite Week, it's not in end of financial year, it's not in Black Friday, Cyber Monday. So, what I find really fascinating is the ability of the model to adapt and respond quicker than a human would in analyzing billions of data sets in, you know, literally milliseconds to be able to respond to all of this. So that was a really interesting case for sure.

SPEAKER_03 21:22

It sounds like your customers don't like positive reinforcement, Johan.

SPEAKER_02 21:26

Well, I guess they do on a BAU day, but not on a salary. I guess that's the interesting thing that I found.

SPEAKER_00 21:32

Yeah. It's because, you know, intent differs. So, you know, your September period, Johan, like it was just lightning in a cup because we'd had the, you know, six to eight weeks to train the models and switch on the causal AI variants that, you know, are more accurate. You had a higher traffic that had high purchase intent. So the model's always looking for purchase intent. How can we influence this visit? So, you know, when you do those things and you can move conversion, and we know how. Much revenue increases when we even move conversion by a couple of percent. That is when the real magic happens. But yeah, it then needs to be able to adapt to a normal period. And you know, all the way throughout the year, there's no urgency. You don't want to use coupons or promos or discounts. That would just erode margin. So we need to leverage, you know, other things that really help customers make a decision when they're shopping on the site. So yeah, we need to be quite adaptive and responsive.

SPEAKER_03 22:28

And when you talk about being adaptive and responsive, you know, you talked about running multiple tests and making the decision of doubling down or stopping or doing nothing. In Bahamix, is it that this is done automatically or do you still need a controller at the wheel to turn on and off different nudges and different bits of functionality?

SPEAKER_00 22:47

Yeah, it's a bit of both. It's a self-learning AI system. So that decision intelligence is happening in real time and in an automated way. Look, once we've been able to hit statistical significance, we can then switch on the causal variants of the nudges, would just improve the effectiveness even more. So it is a human to AI partnering for sure. And you know, the humans are building out these neural networks and working. But yeah, the AI is just put on autopilot because it can respond much faster than a data science team or a trade team can. So yeah, that's the combination.

SPEAKER_03 23:23

Doing much tinkering, Johan, or are you letting the bots do their thing?

SPEAKER_02 23:26

Uh, this is definitely something uh I'll leave to the bots. It's very fascinating to follow, really. I mean, from what we do on our end, potentially then suggest different types of nudges or different design of the nudges, and then you know, let AI test that.

SPEAKER_03 23:41

Is there a nudge that you've implemented? Or, Brooke, I'm keen to hear about the wide library of nudges that you've implemented with other customers that you were like, I don't think that would do anything significant. Like when you talked about having this person just purchased and having the 50% uplift or 17 people just purchased, I'm like, I always look at them and go, does that really do anything? But obviously it does something to our brains. Brooke, I'm really keen. Outside of Elite Sups, has there been any nudges? I know you've got a library full of nudges that you're trialing. Any that have really surprised you, any that have kind of knocked your socks off in terms of, I don't think they're gonna do anything, but actually had a huge impact.

SPEAKER_00 24:19

I mean, even with elite subs, right? It's how do we even craft around this goals-based shopping behavior when you're, you know, shopping for supplements? So even when customers are facing hesitation, leveraging things like cognitive dissonance once they get to the cart. So, you know, they might have gone through the heavy lifting of choosing something, getting to cart, and then the hesitation sets in. So just putting in front of them, you know, your PB just got closer.

SPEAKER_01 24:46

Oh, that's it.

SPEAKER_00 24:46

Or, you know, things like that. So, you know, now in your cart fueling your next session. So that's the endowment effect. I feel like I'm already taking it, you know, and I I can see how it would fuel my next session. So just to get over those last little cart hurdles and some of the interesting ones we've deployed as well, is even if a customer, I guess their intent is starting to waver, just changing the Chrome tab description. So they might have already popped onto a competitor site or gotten distracted and gone and done their email or something like that, just saying, you know, your nutrition awaits, or you know, just bringing them back into this is why I was there and making that decision in the first place. So, you know, yeah, it doesn't need to be sell, sell, sell all the time. It's more about reframing for how the brain works so that we can help the customer, but just not lose the sale because it's when they're on the site. It's when they've been live in the session. You know, we we don't want to have to run after them after they've gone down the road.

SPEAKER_03 25:44

I did notice that, Johan. You do such a great job of the what's in it for me from a customer perspective. The first thing you're hit with when you enter your site is what's your training goal. So it's not like do you want protein, do you want collagen, do you want whatever it is? It's like, are you trying to lose weight, build muscle, endurance, whatever it is? It's a really interesting entry point.

SPEAKER_02 26:03

Yeah, I mean, when we change that, so basically I guess we turn the navigation into a first-party uh preference data that we then, you know, that data we can use like everywhere. The first we actually use it for is in uh our email automation flows. So someone selecting, you know, build your muscle as a goal, then you know, that whole welcome journey will be around that and the products that fits into that category. The other thing this did for us is it created a new segmentation layer for us to use in promotion and you know, with suppliers and whatnot. And then I guess the long-term goal that we set up when we launched this is to be able to use these signals basically in our whole tech stack. And you know, that includes Behemoths to feed these underlying signals into nodges so they become even more relevant and uh you know that resonates even more with uh the consumer in that specific time. It could be as simple as you know, not saying, Yeah, this is great for weight loss, something like that, if we know that the customer is on that weight loss journey.

SPEAKER_03 27:07

And you mentioned that having those segments helps you with suppliers. How do you use that information with suppliers?

How to Convert First-Time Visitors

SPEAKER_02 27:13

I guess we kind of articulate that we have the ability to segment on goals and you know your your product fits this type of goal, and you know, we have X many people in this segment that we can, you know, leverage.

SPEAKER_03 27:25

Yeah. When I was looking at this, and it makes a lot more sense now that I've heard about Elite Week, which is one of the best promo names for a week that I've ever heard. But I was thinking of it through a lens of I think what a lot of retailers are struggling with at the moment is new customers because it seems like there's never been a time where more people are talking about having goals for their e-com business where they're just trying to hit a certain amount of new customers and then assuming that the revenue and the basket size will take care of it from there. But new customers is a metric that seems to be at the forefront for a lot of people right now. Brooke, I'll start with you. How are people using behavioral science to help convert that first-time visitor? Because we're spending a lot of money on Meta and other platforms to attract them. How are you finding that the best brands are using behavioral science to then convert them on site?

SPEAKER_00 28:17

Yeah, that's why we started Nudge. You know, so much money is spent getting the traffic to the site and yet Cartivadamant sits at 70%, you know, even despite all the payment changes and checkout changes. So yeah, what what can we do to nudge a customer when they're a first-time visitor? You know, they might need a bit more reassurance and trust, flags, and and symbols. Whereas, you know, someone returning that's either familiar with your brand or product category might need, you know, other things to help them make a decision, informational nudging or maybe pricing nudging or you know, so yeah, we can deploy different types of nudging for first time returning visits, but I actually kind of find that a bit too black and white. You know, it's not just first time returning. It's yeah, what type of site do you have? Are you a multi-brand, multi-category retailer? Because, you know, if you are, there could be a different decision lens taken every time the customer comes back. Are you a marketplace? Are there different shipping options and you know things that are maybe wavering trust? And then, you know, uh add to that, I don't know, are you buying a t-shirt? Are you buying a vitamin for your pet? Are you buying a dishwasher? They all have very different decision-making journeys and therefore the framings that we need to take with behavioral science. And then you just add the human element like are you scrolling on the bus? Are you cooking dinner? Like for me, if I'm trying to sit down and do something at the computer, it's distracting me like every single time. Or I don't know, if you're investing a lot in TikTok, literally that's like your brain's hardwiring was to go back to that dopamine hit in the feed.

Understanding Customer Behaviour

SPEAKER_03 29:48

Your customers aren't coming on to read a PDF file, are they?

SPEAKER_00 29:52

No, they are. You're finding that, you know, I want to go back to that feed and get that hit. So, you know, we need to be able to respond to multiple inputs and when the customer is there because you know, maybe what they did last time is completely different to how they're showing up today. And our role is just to help them make a better decision and to drive that conversion for the retailer.

SPEAKER_03 30:14

And Johan, what have you found in terms of what works for you in converting those first-time users? Is there anything that's really shone for you?

SPEAKER_02 30:21

We can't really uh do the, you know, pull the price or discount lever because uh we have a store network that we need to protect. So I think for us, we need to build it around trust, relevancy, and you know, we're here to help. We are the experts. And, you know, having someone like Bamix where we can increase conversion rate and average order without pulling that discount lever, that is gold for us. Because I can't go out and run discounts all the time, even if I wanted to. Yeah. Then I would have the whole store network after me.

SPEAKER_03 30:54

And half of them are your stores anyway, so you'll be after yourself.

SPEAKER_00 30:57

Yeah, you'll be chasing yourself, like, ah, two zones. I guess like in e-commerce, it's not enough just to, you know, predict, okay, is someone gonna abandon their cart. We need to know why. Yeah. We need to be able to explain the cause and the why, because if you just slap a promotion on everything the minute, you know, the customer's about to abandon, that's going to erode your margin. So knowing why helps us show the right nudge at the right time. And you know, what I see with your data, Johan, is even when we're using the social proof, which is building trust and is quite strong for a first-time visitor, they're actually shown. It's interesting to see how it's shown with the cognitive dissonance or, you know, in your sale period, magnitude congruence on the mobile phone, being able to see the discount smaller because that's really fluent for our brain. So yeah, when when we lift the hood, it's like, how do they all play together so that yeah, you don't need to put a coupon on top.

SPEAKER_02 31:50

Yeah. And I mean, we do tons of A-B testing on our P2P's and checkout, but there are there are so many variables that you know it that you can't discover that is hidden in the data that you know you you don't understand.

SPEAKER_00 32:04

Yeah, and we don't want them always to be like a pop-up or a fly out because again, that gets a bit or this is marketing. We're wanting to frame the decision. So sometimes it's just changing a color or you know, framing a default of what size of the product you should buy to stock up, or you know, maybe it's intelligently sorting the items in the cart to anchor them to the most optimal conversion. So, you know, a customer that is not in your face, it's a very subtle, we call it a silent nudge, but it still helps the customer, you know, on their journey and helps optimize a bigger revenue without having to put the promotion in there.

SPEAKER_02 32:39

Yeah, and I think for for supplements as well, I think first time customer comes in, we have so many different types of supplements, and then you have so many different types of protein, and they know they need a protein, but then you know, they're looking at one protein and then they see another one as like, okay, I know I want one of them, but I I'm not really sure which one.

SPEAKER_03 32:57

Yeah.

SPEAKER_02 32:57

You know, throwing in a nudge there just to boost the confidence or reassurance or whatever, that that would help from them bouncing that session.

SPEAKER_03 33:06

And how do you determine what nudge is right? Because obviously, in those situations, you could nudge someone towards a goal-based recommendation based on the product, or you could but nudge towards something that might be a little bit friendlier from a GP perspective. How do you balance those nudges between the commercial outcomes? Because we've seen a lot of this with personalized product recommendations and search and merch tools. How do you find that balance between nudging customers in the right direction to meet their outcomes versus pushing them towards products that may have a bigger commercial outcome for you in that transaction?

SPEAKER_00 33:41

Yeah, Nathan, I guess you're hitting upon attention, right? Because we want to drive conversion. That's the goal, but not at all costs. You know? We have to be truly helpful. And in the world, there's nudges and sludges. I think Sonia might have even talked about that on her episode. And that's the dirty dark pattern area of nudges, which we really need to avoid. Look, some examples even outside of supplements is I think obviously the promotion one, we don't always need to put a promotion on it to drive the sale. But even things like if our model has detected you've added items to cart, you're probably going to return anyway. We intervene at that point to say, hey, do you really want to order this item or do you really need this many sizes? And we deploy it intelligently using things like loss aversion to really get the customer just to stop in their tracks. And that, you know, if we can reduce returns, that's going to improve the overall margin. You don't need to pay for the return, bring it back into warehouse. So yeah, we don't want to drive conversions at all costs because we know for retailers in the warehousing and the shipping costs, there's a true downstream impact to that.

SPEAKER_03 34:46

Yeah. And there's a trust impact too, isn't there?

SPEAKER_00 34:48

Yeah.

SPEAKER_03 34:48

Johan, in terms of you mentioned their curveo and you mentioned that you're running on Shopify. I know you've got a lot of tools at play here, and you mentioned the myriad of A B variables or A-B testing that you've got going on. I think if anyone's listening to this and they're like, look, this sounds really interesting. And I would love to have this kind of personalization and science behind understanding our customers. There would also be that tension of like, oh, it's another tool on top of what we've already got. How do you balance this out? And how do you ensure that something like a Bahamix is adding increased personalization on top or with the tools that you've got there rather than cannibalizing what you've got?

Integrating Technology for Enhanced Customer Experience

SPEAKER_02 35:30

I think with all tools, you you gotta make sure they don't interfere in the same session with the with that specific customer journey. They should support each other though. And I that's what I think we have here. So, you know, with Cove, we had uh the search and discovery. So you guide the customer to the right product. And then you have Behemoths with nudges that you know reduce that hesitation and build the trust. And then, you know, on top of that, then you have Clavio, which is the life cycle and it keeps the relationship going. And then we have the PIM, which is our uh social truth when it comes to product data. So, yes, we have a lot of softwares, but it creates one experience and they are all supporting each other. And we have had in the past, we've had systems that kind of interfere with each other, but the important thing here is to find uh eliminate that and make everyone support each other rather than compete.

SPEAKER_03 36:23

I love that view because it sounds like each is not necessarily generating data, but taking data from other platforms and building on it or using it in new ways. Like I love where you talked about having new segments that you can use for customer experiences or supplier information. And I can use, assume it's used in multiple other ways, but it seems like there's a bit of a flow here to just one big great data experience.

SPEAKER_02 36:46

Yeah. I mean, surely you would allow to have one software doing everything, but you're not really our data.

SPEAKER_03 36:52

Yeah. Brooke, how do you normally have those conversations? Because I I know there is that tech overwhelm at the moment. How do you show people that this is actually using data that they've already got and then enhancing it and delivering a new experience that they couldn't get elsewhere?

SPEAKER_00 37:06

Yeah, I think what Johan said was really key. It's you know, how do all your tech platforms complement each other? You know, something like a search and merch that's help me find what I'm looking for. You know, where we step in is just help me decide.

SPEAKER_01 37:18

Yeah.

SPEAKER_00 37:18

So it's, you know, how can the the tech platforms complement each other? And I guess what normally comes out in most rooms that we're in is it's incremental revenue. How can you drive revenue on top of what I'm already doing that isn't, you know, competing with with something I'm already delivering elsewhere? And generally, you know, we're looking at things that are invisible to most of your e-commerce dashboards. You know, we're mostly looking at the funnel, where's traffic coming, where are people dropping off conversion rates, but nothing really explains it why. So, you know, looking at something that knows hesitation, lack of trust, and can just deliver what you're leaving on the table because of that, that's generally where the most powerful conversations happen for us.

SPEAKER_03 38:01

I think that's an interesting piece too, isn't it? Because no one likes doing reports or understanding behavior that has happened. But if you're analyzing behavior to do predictive behavior, that's where it becomes interesting because you're then picking up those opportunities.

SPEAKER_00 38:15

Yeah, I guess it's one step forward than that. It's we're not just predicting, we know why, and we're gonna intervene in real time. So if we just take a data set and we look at what they did last time to predict how they might behave today, that that is just not a true indicator. Statistically speaking, like customers just change every day that they come to your site. So yeah, in in real time is really where we find the sweet spot of revenue.

SPEAKER_03 38:40

And I've got to ask about implementation, Johan. How simple, easy hair pulling was it?

The Simplicity of Implementation and Data Utilization

SPEAKER_02 38:45

Honestly, it was probably uh the simplest and easiest implementation uh we ever done in terms of you know the technical side, because it it basically one line of code that you just need to add. So anyone in the team could have done that.

SPEAKER_03 38:57

Seriously, one line of code. We hear that too often.

SPEAKER_02 39:00

Yeah, and uh in this case, it is true. I guess from there, you will have straight sessions, mapping out the journeys and and see what type of nudges you want to deploy and how will they uh look so they they feel the same as your site. And so that's a bit old planning, but you know, overall it's very simple and easy, uh straightforward implementation process.

SPEAKER_03 39:21

And Brooke, in terms of looking at customers who would fit for Bamix, is there a sweet spot? Is there a certain tech stack? Is there a certain revenue number? How do you know who's right? Because I assume you've got to have a base level of data.

AI's Role in Personalization and Customer Engagement

SPEAKER_00 39:35

Yeah, we do. Yeah, you need a base level of data for AI. It's a self-learning system, but we work with anyone from you know 50,000 visits per month up to millions of visits per month. So and look, I guess we're all human. So retail, we work a lot in retail across yes, supplements, that fashion, furniture, uh electronics, but even into spaces you wouldn't think, like buying a car with Volvo or insurance, picking the right insurance. Like these are actually bigger decision-making journeys than buying a t-shirt. So yeah, we just start with the science and and we work from there.

SPEAKER_03 40:09

That'd be fascinating getting into people's head as they're buying insurance or cars.

SPEAKER_00 40:13

Well, it's quite cognitively taxing, which is why I think we put something like insurance off.

SPEAKER_03 40:18

Yeah, I couldn't imagine it. That's very cool. We have mentioned AI, but Bahamix kind of isn't riding the AI curve as I see it. Like it's been built with AI, you know what I mean? Like it's it's been a data platform as I see it. Brooke, you might tell me totally differently, but it's not like you're riding the wave of AI. It's fundamentally part of what you're doing. How do you approach AI and in terms of the capabilities that are opening up? How are we going to get smarter with behavioral science with AI?

SPEAKER_00 40:48

Yes, yeah. We're not quite riding the wave because we were before the wave.

SPEAKER_01 40:52

Yeah.

SPEAKER_00 40:52

So we were founded before these foundation models, the the LLMs like OpenAI and Anthropic were out in the market. So we've built our own AI and trained it on our own systems and service from the ground up. So yeah, I guess it's, you know, rather than riding the wave being AI first, it's about well, evolving what we've done to make it smarter and smarter. But then looking at, well, how do we pull in Gen AI or LLMs where it's needed? So we do sometimes use them. Like if we want to summarize your PDP content in a more salient manner, we can use, you know, Gen AI to shortcut that. But yeah, the way we see Gen AI is it's amazing. It's like, I don't know, it's like the fastest car that can get you anywhere. But then we see causal AI as like the map that's like, okay, if you turn left here, you're gonna get to exactly where you need to be with the result and outcome you need. So yeah, for us, it's about using the different tech in a way that's just gonna ladder up to the goals and strategy.

SPEAKER_03 41:50

Makes sense. Johan, how are you saying AI at the moment, just in general, like less so behavioral science? But are you a fast adopter?

SPEAKER_02 41:58

Yeah, but uh more on uh we're trying to use AI where uh take away the boring elements of uh the job and put that time into where we can make a difference. And I think AI does a you know really great job in you know the personalization of creating those experiences, which uh human can't. But for us, you know, it's it's really important because our identity is we genuinely want to help people. And I would never get AI to take that responsibility. Uh I'm sure we're gonna, you know, with the Gentle Geek uh commerce coming, uh they're probably gonna be a part of it. But with our staff are so passionate and you know, we really, really care about the customer's fitness journey. Yeah, that side of the company will, you know, we want to keep that yeah, we want to keep that human. But optimization, personalization, all that AI 100%.

SPEAKER_03 42:48

Fantastic. Now, the next 12 months for Elite Subs, what is on the radar? Obviously, more stores, I guess. Uh, yeah, most more stores. International domination.

SPEAKER_02 42:59

That would be nice. I mean, yeah, it's it's definitely um, you know, we we are expanding in Singapore, which is um interesting market for us. And you know, now with Dubai and New Zealand coming on board, we have some other uh really interesting projects that I hopefully can come back and talk about at a later stage. But in the context of this conversation, I guess you know, we need to stay relevant. And with that, I mean, working on our product data. So again, with AI, you need quality product data to you know feed in and or signals, in this case, our goals, for example. It doesn't matter which, you know, if you have a super AI, if you if your data is, you know, at best average, you're gonna get average results in the end. So that is something we're working really, really uh hard on this year. And then with Behemoths particularly, it's gonna be you know, try and leverage those underlying signals a bit more.

SPEAKER_03 43:52

Love it. Now, Brooke, I can't imagine the world of behavioral science slows down at all, especially with the explosions in AI that we're Given that you are ahead of the wave and not on the wave, what's next for the hyme?

SPEAKER_00 44:05

Oh, yeah, actually just signaled it. We're pretty excited about it. Yeah, we're now looking at it. We have amazing underlying data based on causality, probability scoring. So being able to feed that into the CRM and app to make that more is where we're taking it next with signals. You know, we talked about the card of it before. Usually you get an email with the last item you left in your card. It actually might have been the other item if you went up to where there might be a higher probability. So we're able to feed that probability story and multiple different types of behavioral data to in the and even to make money on it. Like is there a customer and item is wasting money on it? So that's really exciting. I think they go really hand in hand. Yeah, we just already today we apply the the causality to things like inventory and merchandise. Uh excited to use SEO. So yeah, we're really about just how do we optimize the whole digital journey.

SPEAKER_03 45:01

I think that's really, really incredible, especially as we know the the journey is being pulled apart. You know, we're gonna find checkouts in LLMs and everywhere around the web. So taking that that expanding that data and that personalization out beyond just that website is gonna be a huge unlock. And I love what you're just saying there. I know it was a simple example, but around the abandoned cart, around, you know, I don't know any retailer who loves sending someone an email four hours after they've had something in their cart that they've abandoned, having to send them a discount to try and get them back when they've already showed intent. So being able to do that unlock, whether it's that right product or not, is really cool. And that that's just one example. So I am super pumped to see what you can do, especially for those ads, because we don't like paying any more in the ads than we need to be doing. So it's about time they got a bit more personalized. Brooke, Johan, thank you so much for joining me on ad to cart. I would like to say that I'm smarter, but I actually feel dumber because Brooke, I'm gonna have to look up some of those words that you talked about to get my behavioral science up to scratch. But amazing to hear the unlocks that you're getting, Johan, and the incredible growth of Elite Sups. And I really look forward to seeing more of what's coming out at Bahamix. Brooke.

SPEAKER_00 46:12

Thanks, Nate. Yeah, great chats.

Bushy's Takeaways

SPEAKER_03 46:14

Yeah, thanks, Nadem. It was great to be here. What an awesome episode. I always, always love nerding out on behavioral science with the Bahamix crew, whether that be Brooke or whether that be Sonia, it's always such a fascinating topic, and I just feel like we can never learn enough about it. What struck me from that conversation is how much of what we call optimization is actually just noise or guesswork and adding more urgency, more pop-ups, more friction can actually piss off our customers without even understanding our customers more. Does that make sense? Brooke and Johan are actually working on the cause and not just trying to jam in as many messages as possible. They're thinking exactly what the customer need at this point. All right, three things that I think are worth taking away from today's conversation. If you are hard-coding nudges site-wide, you're probably training your customers to ignore you. Showing 17 people viewing this on every single product, on every single visit, it's not social proof anymore. It becomes wallpaper. Customers see through it very fast, and once they do, it starts costing you trust. Dynamic nudges, shown to the right person at the right moment, on the right product, for the right experience, are a completely different thing to blanket urgency tactics. Secondly, understand why someone is hesitating before you intervene. A near abandon driven by decision fatigue needs a different response to one driven by a lack of trust or a bit of distraction. Slapping a discount on every cart that's about to go cold solves the symptom, but it erodes your margin and doesn't actually fix the cause. Knowing the why is what lets you show the right nudge, not just any nudge, and make more sales at full margin. And thirdly, every tool in your stack should own a distinct job, but they should all be building towards the same picture of customer intent. The way Johan framed it landed really well for me. Coveo finds the product, Bahamix reduces the hesitation, Clavio holds the relationship. And underpinning all of it is a navigation decision. Asking customers, what's your training goal? Instead of what product do you want? That single UX shift created a segmentation layer that now feeds email flows, supplier conversations, and behavioral nudges all at once across all interconnected platforms. Understanding intent is the foundation here. Everything else builds on it. If this episode's got you actually thinking about what's going on in your customer's head and what you might be leaving on the table because of it, come and continue that conversation over in the AdDicart community. We would love to see you there. There are over 600 e-commerce professionals sharing tips, questions, and all sorts of things, maybe getting inside customers' head as well. Join us at adducart.com.au. We would love to see you there. It's free to join. Thanks again to Brooke and Johan for a really fascinating hour talking about behavioral science and how we can use that in e commerce. If you haven't already, hit subscribe wherever you are listening, and we'll see you on the next one.