What does good data infrastructure look like for your businesses or more generally? Well, what it looks like or what is at our business, because... Well both, well both. Two a worlds apart. Yeah, yeah, yeah. Tell me about both. No, so basically, I like to say that we have a very challenging tech stack with data for legacy reasons. The way that Flying Tiger grew was through JV partnership. So every single country has a different ERP system, has a different POS system so end of the day, we don't have a data lake or easy access to data. So that's what it's look like. It's a nightmare. What good looks like, I think everybody knows, then we go for data lake, big queries, single view of customers, all of that. But I have to say that we are miles away, so then we can discuss it, what are we doing to overcome that, but yeah. How do you feel about the overcoming of that? How close are you in terms of the steps that you're maybe taking? So I think we are doing the right thing. I don't personally believe in tech revolutions or big reorganizations. So the way that we did with, I don't know if you guys are familiar with the Mach Alliance, but it's basically creating APIs and ecosystem outside of the complexity of the data that we have, to be able to slowly erase our data and migrate and not slowing us down. I think a good example is the Flying Tiger Club that we launched a few months ago. And we built it entirely outside of our ecosystem and we found a way to connect to the ecosystem only on what we needed to match transactions and things like that. So building those connectors, it's the way that we found and it's working pretty well. It's flexible, it's scalable. And it's most importantly super cheap. I mean, super cheap, it's cheaper. Over to you in terms of data / infrastructure? Yeah, I've probably got a similar complex landscape like Andre's outlined. So in our business we're covering UK and Europe, again two systems. In the new world, in the very near new world we are changing the ERP systems so that's great and gives us one step closer to unifying the data. We rely on that unification really, and it's viewing the data from so many different points of view, from what the customer's buying, let's say the merchandise and the sales analysis, what we can see across the sales channels, and then certainly from a marketing point of view the customer. In our business, we have two types of customer. Our main customer is a business customer, it's hairdressers, it's beauty techs, barbers and you know all of those kind of customers and that's a real key growth for us and because of that we trace every purchase so we've got a great deal of insight and data and can see the journeys for a business, a B2B customer. On B2C it's growth for our business and we're trying to put the steps in to recognize all the transactions and to understand those journeys. So very similar, in March we introduced a loyalty program that resets it for a B2B customer, but actually for a B2C to really build on that recognisability and the contactability. And then, fine, so you've got all these things that is like, the information that you've got at your disposal, some of it more fragmented, but with great scope to hopefully do lots with it. How do you then balance speed of execution? Ideally, you do it at pace with having to follow, we talked a bit about this, various regulations or limitation that comes with having that data. And then also making sure it's tailored to your customers and the way you use it. So speed, personalization, regulation, there's so many things to think about. Where do you even start? The main reason behind the Club, that's how we call it, is to be able to have the first party data. With the new regulations, it's very hard to map customer shopping behavior with all that. So, and that is a great example, the way that we treat things. It's very speed, customer first, and then if it doesn't hurt us, we go out there and we test and learn. So that's how we enhance the data. That's how map the customer journey. That's how we make decisions is by, for lack of a better word, creating MVP products, putting out there, testing and learning. So very simple products and then we can either trash or build on top of that. So that is also a way to overcome the limitations on data. And of course, the golden word for us, it's assumptions. So if you have well, two variables, you can conclude on 20 other variables, but yeah, those together is how we do it. But the most important thing is to put something together in front of the customers and then change it, because any survey, any data, et cetera, doesn't tell you the naked truth, how the customer would interact, what they like, what they don't like, and especially how they are changing, because they are changing all the time. So I'll say that's how we do it, fast, speed to market. Learn fast, fail fast, and then it's scale fast. So that's how we do it. Yeah, I mean, our key barometer is sales. We want the sales from, we want the sales conversion, I mean, I think that's why we're all here, sort of thing. So from our point of view, we'll test and learn through the different channels. We have 450 stores, we have a growing e-commerce business and we have actually a Sales force as well. So we'll test and learn in terms of those, the go-to-market planning, the project releases and so on. But yeah, that's our main driver for that. And then we'll take the insights of the behavioral, the Customer and where they shop in on the back of that. Um, and I, this is why I'm going to bring AI into the conversation a little bit. Uh, I think a lot of the businesses were using AI before it became a thing, but now we're talking about Gen AI. It looks different, is smarter, lack a better word, but people usually had some form of machine learning in their business way before all of this kicked off, but it's moving very quickly and I feel like it is opening up a world of opportunity, but at the same time, you're not quite sure where to look. I don't know, you might tell me different. So I'm interested in hearing from you in terms of how do you think about it, maybe use cases or not. You know, try not to get distracted by it. So, yes. Yeah, I mean I think we talked about it on the call, though it's evolved and it continues to move, but we are focused first most in the machine learning and that improving execution. So working with international markets, translations is a really big thing for us, so using AI to speed up and accelerate communication to market has been key. Going forward we're using it in a far more sophisticated way. Key things have been around unifying the data and the views across the various different data sources and also really with customer segmentation and targeting in those audiences we're much faster on those segmentations. We use systems that give us a propensity to spend what would be the right audience and that just speeds it up because as much as we've got an analysis team it's a much slower process. In terms of, you obviously mentioned unifying, how easy or difficult has that been? Do you just plug the stuff in from different teams and, you know, does it bring it all together or? Yeah, it's hard to get, I mean, we're working and trying to move to a place where we'll transition to a CDP, customer data platform, and that will make it easier. But it's hard, I think, to identify where it's actually coming from, from an incremental point of view. You know, you'll have a merchandising team or a buying team that will think it's about the product. Obviously, with marketing, I'm going to think it is about whatever marketing they've done and so on. Could be the sales team, it could be that they like the web banner, we'll never know from that point of view. So we look at it again through a sales lens, but where we are really interested in is how that correlates and how that unifies against the customer piece. So, we will use the data in that web through, okay, they're bought in this way. And what we tend to do is look at over a longer period of time to look at that buying behavior particularly in what we're selling and the frequency of shop is often quite low. Yep, AI... So that's a big topic. First I will explain how we try to implement and then fail. There was from the top down, trying to do a gigantic project of AI and optimize all the projects and connect all the dots. And once we probably finalized the first PowerPoint, putting together the project was a rather obsolete because AI changes every day. So, then, after a few months... reflecting, discussing, we decided on an approach that I strongly believe that works and it is working very well, that is a bottom-up approach. It's oversimplifying AI. AI can be divided in many things, like AI could be compared to internet. So you cannot tell functions what to do with AI. You need to do a proper training, enable the platforms, give them the platforms and let them use it. I think that is the first step. And then control the outputs. I like to say that we have a copy and paste approach so it doesn't mean that I can get, it means that I need to get a data from one place, understand, do a prompt on another and maybe utilize it another to generate an image, an asset or whatever. So it's not connected, but that's how we found a way to do it fast and bottom up and empower everyone to use it AI. An AI in marketing can be used to create assets, can be used to segment, can be used creating is not really AI, but can be used to create communication, but also can get one step further that can create assets with the right colors or the right fonts or the images to be able to track the performance of that. So you put it on the back end and then you can start tracking the performance so on top of views, engagement, shares, clicks, whatever, you also have the image – so we start having another source of data. So basically that's what we did. It's let everybody use it. It's super cheap, the license to use it, control the output, be comfortable with mistakes, because when we empower people, some mistakes will happen. And then that creates a learning cycle. So I think that is the easiest way to do it. But then outside of marketing, I mean, Flying Tiger, I don't think anyone will do a business plan or market research or industry analysis without utilizing the Chat GPT reasoning model. It can be done in seconds, otherwise it would be ages to do, right? Or finance or legal to revise contracts. So a lot of things. And try to, of course, comply with every regulation, but don't be too precious with the data that you're sharing. Yeah. Because on the end of the day, if you are too precious, I mean, all the AI is built to get your data, right? It is their business model. As more data you put it, more they incentivise you to use it. Otherwise, they will not. So, not be too pressures on that, otherwise, you don't get the most of it. And of course, Flying Tiger is a small company, so I don't think anyone is trying to look at and hack what we are doing so there's also a positive side. And I think there's the third layer that is then the full scale of AI, and then I don't know how that's going to develop. It's when you connect the dots, right? When you connect to your data, results with a generative tool that creates an asset. So kind of the entire funnel or the entire process. But I think that it's, in my view, it's gonna be very hard to do it because it changes all the time. And it's going to be, the output is going to as good as your data. And I don't know about you guys, but data is always a challenge in the company, right? Doesn't matter what type of data. If it is product descriptions, if it is customer email, if it's golden keys, whatever it is, it's all a tricky, so. I believe that I always want to have to be a human interface in the AI, for other things not, like translations, etc. Yeah. Then you can just create a tool that it learned by itself, etc., and then gets better and better and so, yeah. That's my view of AI. I keep I mean, it's my own view of it. I keep hearing this a lot which is yes from companies I speak to yes it is useful but we're at a stage where and possibly for a long time oversight like human oversight has to be part of it and it's what determines how you ultimately use it and what obviously the final output is you can't just plug it in let it do its thing and hope for the best. So it feels like that's where we are, but also... hopefully, I don't know, in a couple of years, we'll move towards what you said, which is the... Maybe, I think we all gonna as, we as a human, we don't change the habits frequently, right? During the pandemic, everybody freak out and say, oh, all the brick and mortar stores are gonna close, nobody's gonna want to shop in the streets. And then as soon as the pandemic ended, everybody went back to the streets, so I think as customers are gonna get sick of that much AI information or that much content. So we're going to have to have some shifts. Of course, AI can learn that as well and then can adapt it. So it's going to be a long game. So if I knew the answer, I'll be working with Sam Altman here. Dara, you obviously mentioned consumer data platforms. Are they, you know, a bit maybe unfair to ask, but do they live up to the promise? Is there a hope that they do? What does a really good one look like? Remember you're on SAP. I think it's a good question, I mean we know that we need to get there and we need to make it easier to get a single customer view but I wouldn't even call it that just a unified view and so certainly from from our business point of view some of that is slowed down because the ERP transitions it is crucial and we're going through that right now. In the meantime what we're to do is sort of like just... get the right metrics, get the right things that measure in and prove the point, whatever that may be. And we've spent a lot of time cross-functionally to really try and establish that and look at it through the same lens. Same, but different but to what Andre was saying about, you know, control and governance around like AI, in a similar way, we've spend a lot of time trying to... unify, work with common prompts for example, and to shape how we say things. Work in hairdressing so it's hair color, well actually a retail customer would maybe call it hair dye and you know at a basic level just building up those views but as a professional they call it hair color so you know really trying to build that so we're all at a common place whenever we use such tools no matter what department or what function you're in. In terms of the actual people element and culture, how easy is it to bring people on board and your teams to sign up to whatever you're doing or hoping to achieve? I think often these transformations, yes, they have to have great tech. Yes, you use the data, but it's also the people and do they buy into whatever the vision is and how difficult is it to achieve that for them to understand what the, you know, this is the big picture, this is why we're doing this for. Culturally, internally with your teams, do they get it? Yes, I would say. I don't think, I mean, they work in marketing, there's a lot of younger people in this business and always attracted to the shiny new thing, so the tools that come with that are equally appealing. Sally Europe is part of a global business and mothership is American and what we found actually in the use and the adoption of some of these tools, you know, there are differences, GDPR differences and so on from overseas. We're trying to navigate governance around that kind of thing, but adoption has been easy. It's just making sure we're working with the right framework moving forward. And it keeps changing and keeps moving. Andre? I would say that the vision and the purpose, it's very easy to engage everyone towards it. But I'd like to say that it's just like drinking water, right? We all know that we have to drink, whatever, three liters of water per day, but nobody has time to drink three liters per day. So on the AI it's the same thing, or on that revolution, as you call it, it is the same. Everybody knows that we need to change the way that we do, but no one has time to stop the test that they are doing right now because it's for yesterday and learn a new way of doing that will be much better for the future. So on that sense, that's what I say, that we need to train and to have right now, and it's interesting because everybody say, oh, hey, AI, everybody's gonna lose job, everybody's gonna go homeless, whatever. But no, right now it's actually increasing a little bit the resources so everybody can start learning how to do the new functions with AI. So it comes into play of optimizing what you are already doing it. So that's why I like the bottom-up. And then after that, we can discuss it about top-down revolutions, et cetera. But now it's more about, I don't know if you guys are familiar with, but mid-journey or whatever. I mean, I come from fashion, so before was like weeks to develop a product. Now it can be done in... seconds, but you need to learn how to do it. But you don't have time to learn how to it because you're doing the old way. So freeing up a little bit of time, giving training, giving access to the tools, and then naturally people will adapt it to the new way of working because everybody's eager to do. And once they experience, it's mind blowing. So, yeah. That's what I've got to say. And then in terms of how, what do you think about, obviously we talked earlier a bit about loyalty and customer experience. When you think of ROI and how do you know, do you monetize or make sure you use this data to hopefully either lead to more loyalty or a better customer experience or ideally, you know good return, which one's... is one that's more important than others? Do they work in unison? How do you think about capital and what that means? Where we're trying to get to and I'd say it's working progress is to use data and be more data driven but to use AI tools in every part of that journey really. There is a functional benefit that drives faster execution and can accelerate messages at whatever point but there's also a key place I think in the insights and really forming and shaping the intent. Um, so it's always moving, right? But I'd say we're more au fait with the bottom up with the execution pieces, but where I guess I'm challenging my team also is to, okay, how do we wed that together? The data, what's that telling us about the customer insight and where then do we use that ahead in our planning strategically? Yeah. Yep, that's where we're at. How long do we have for that answer? A couple of minutes and then we'll open it up. I'd just like to say that I'm not a strong believer in loyalty. It doesn't mean that we don't value the have bias, etc. It means that we all know everyone that buys at some place and has a member card or it's a loyalty, it expects something in return. Nobody goes to Flying Tiger, for happens doesn't find a mug and say I'm not going to drink coffee because I could not buy a mug at Flying Tiger. They're going to go to another store. That is a long answer to say that for me AI it's empowering or not only AI, data it's empowering the way that we see in market that it's reach as many potential customers as we can as frequent as we can, with a simple and meaningful message. So with that having data, then the first two, it's about return over investment, it is about etc etc etc. The last one is where the data comes and the shopping behavior and the AI comes. How can we transform the message into a meaningful message? Because if I open my email now, I probably have like 12 emails with someone trying to sell me something, right? So how do we grab the attention? So that's how I see it. Of course we have the club, of course, we treat our club members lovely, but it's that thing that we use for. So we need to even reach into same customers all the time, so trying to get their mind share or share of mind. So whenever they have something that they can buy at Flying Tiger, we are lucky if they remember us. Oh, you've touched a little bit on this already, but I'm curious if going back to real life examples where you define that data has basically led to better customer experience, or we could talk a bit more about customer experience. Whether that's, I don't know, people in the shop or your B2B customers. So from both perspectives, I don't know, what you've been doing recently around that customer experience? So most recently we've introduced a loyalty program for the Retail customer. We have it for the Trade, for the professionals, but what we were seeing is a lot of retail customers coming in and shopping with us, but we weren't recognizing them. So that's been super useful to really look at the data, where they are, how they're shopping. What the frequency? It's still very much early days. You know, the programs were launched in March. So yeah, we're in the midst of like a 100 day analysis at the moment, but nonetheless, trying to understand the intent. Is that becoming, for example, from your hairdresser recommending the product? Is it, you know, from Google or whatever, from the shopping and so on? Is it from the social? And indeed, once you're in this store, what was your intent? Well, what have we converted you into? Or online as well, so we're in the process of trying to use that to really reinforce the proposition and the offers that we then present to that customer base. I think it's quite a big theme, loyalty schemes, apps. My counter to that is always, is there a risk of customer fatigue because everyone's doing it? And how do you break through and make sure that yours is the one that they keep or they use most often before? Absolutely, I mean, we'll all say that from a loyalty program there's a degree of personalisation that comes with that. In beauty I'm not necessarily sure if that's a good or a bad thing. You know, we want to encourage the customer to shop again and to buy whatever brand it may be, but equally it's quite a fickle sector and I'm sure in your own personal lives you've tried something and then been attracted to something else and so on. So actually that nature of wanting to test or impulse buy, I actually, we want to foster and go on and more. We just want to make sure we're the home for that. On customer experience, I don't know if it's more store-based for you, I think it's both. I think the way we powered data to make the decision, the latest one was to enable click and collect, because one of the main frustrations that the customers share with us was either they go to the store, they don't find it, or they go online and because of shipping costs or free shipping threshold, they could not buy only one product. They could, but they would not buy one product, So, that unify commerce of having online and offline so that they can check before they go to the store if there's the product that they want they can reserve or buy in and collecting the store or then buy on their website and receive at home if they don't want to carry 12 plates around London. So offering, that's what I call the frictionless experience. So basically they can utilize us the way that they need it. So and times they want to go to this store to get inspired but sometimes they go for a specific purpose. I need to, like men buying underwear – it's like a mission. I need go in and out. They need to buy, I don't know, buy decoration for the party that they forgot. So making sure that we cover all of those, that's where we utilize data and then we put it into tasks, as I say, and then if it does work, then we escalate it. So a good example is that Click and Collect, we did a very MVP here at the Tottenham Court Road store. Written in the paper as MVP as you can get and it was very successful. And then it was like, okay, now it's time to have some handheld devices. It's time improve it. It's to turn the store into a warehouse that the staff can locate the products easier. So I think that's one of the things that data told us and was really, really successful. And then on the app thing or the club loyalty, we need to make it as simple as possible. So, I mean, if they have to open the app and scroll down or do a game or whatever, nobody has time anymore, right? So it needs to be inside, I don't know, Apple wallets, Google wallets, super easy, super friendly and have some rewards, et cetera. Of course they have users that want more and engage more, then we also have that. So we need to offer both solutions, the one that wants a very simple solution, the others that want to play a game and earn points. So that's the way how I see it, but it is a tough battlefield. On the, I think, inspiration point, like you said, mission-driven, I say it's easier. You know, you want to, you need to have a haircut. Don't know if that's mission, I'd say so. But, or you know, you've run out of shampoo or you need buy your party stuff, fine. It's almost like you go in and out. When it comes to inspiring people or getting them to, buy more, buy extra, how much leeway is there in, you know, how you market and what you do with your data to inspire them? How easy is it to achieve that? I think we heard, when we were talking about with Steve, you know cost of living, it feels like people are a lot more aware of how they spend their money, what they spend the money on, they might be saving but they're not spending as much. So that inspiration element. How do you achieve that? Does it actually convert? Yeah, I mean, still, I think it always will be. There's a lot of, we have a lot of automated programs, like from a CRM point of view, some of our best performing is abandoned basket, is have you tried, would you like to try, you know, those sorts of things, and they convert. Where I think my, not my recent, my ongoing challenge is, is in a store environment, where it's, I would say it's probably easy for our business to bring that to life. Our store teams, over 60% of them are professionals. So if you're going into a store, you will get some great advice and also the product learning that the teams have to go through. We're a multi-branded business, we sell 350 different hair and beauty brands so, it's quite intense. So they do know their stuff and people seek out that audience. And to your point then, the inspiration or the impulse or converting to something else is is much easier. We're struggling or where we're challenged is bringing that across online. How do we bring the same things to life? Great things like, you know, UGC, EGCI, you know all of those play a part in it. When you're making a choice between, I don't know, wide-plated, straighteners versus a narrow one, you know, kind of need to know your stuff and packaging alone won't, or the description on the listing won't necessarily get you there. So we're really trying to embellish and extend those offline experiences online for that matter. And it sounds like, you know, possibly a lot of opportunity with everything evolving so quickly, where you might be able to emulate or manage to move across that expertise in real time to give the shopper the confidence to make the purchase, hopefully. Yeah, again, like using the data to get to know what to go after and where it needs that relevance and things are just new and it's the latest thing and that's great. Other ones needs to have more relevance or a more authentic voice, so from a brand of communications and our content we're shaking it based on those news. So yeah, on the inspirational point or getting them to buy more? Yeah, for us it's, one, it's creation. So putting together ways that you can use the product together with other products that we sell. So train that content, if it is a picnic, if it's a party, if is a Halloween decoration or whatever. So create that content that inspire people, but that needs to be linked with the sales channel right away. So before we had like the inspiration page, and then the sales page, it's on the same page, but it was a different landing page. Now it needs to be together. A little bit like IKEA does, right? If you go to IKEA, you can see that the room and you can just click on the hotspot and buy whatever you want. So that is one thing. I think another thing that we do on that sense is organizing the store layout or the product display, the panogram, according to cross-selling as well, but also storytelling. So if you go to a Flying Tiger, it's not a commodity store. You're gonna go into a campaign and let's say that it's outdoor so there's everything that it is outdoor right there. So that's how we utilize data to cross selling and to increase sales. But also then feeding back to the product team into the design team, which worked, which did not work, what did sell with what. So they can adapt it for the next collection. So yeah, that has been working well for us. Great, I think we're out of time. Thank you so much, Andre, thank you so much, Dara, I really enjoy that. So yes, let's give them a round of applause, thank you.