Driving E-Commerce Growth Through Technology | Dmitriy Anderson

Justin Aronstein (00:03) Welcome to season two of Check In to Check Out. Today on the show we have Dimitri Anderson, CIO and Head of E-Commerce and Marketplace Strategy at Leroy Merlin. That is a really broad role. So Dimitri, what does that entail? Dmitriy Anderson (00:19) You did. ⁓ Yeah, as we got to discuss a little bit before the recording, it's a hybrid role and I'm very, very lucky to... ⁓ to have it and to be the sculptor of my role really and having awesome CEOs that trusted me with this role. what it really entails is weaving technology and business ⁓ priorities together in order to achieve hyper growth, if you want it in one sentence. Justin Aronstein (00:56) love it. ⁓ I love it mainly because that's the background that I come from. I I went to school to be a software developer and then realized that I have a passion for business. So being able to combine my technology acumen to drive growth and shortcut a lot of things because I'm a technologist and not having to ask another set of teams or not having to or knowing what questions to ask really Dmitriy Anderson (00:57) Thank Justin Aronstein (01:25) drives something that's more interesting and gets you to grow faster. So I'm sure that's what you find as well. Dmitriy Anderson (01:34) Yeah, exactly. And on top of that, it allows you to be knowledgeable enough for marketing agencies to not bullshit you with different fairy tales and stories. You can suss it out immediately and ⁓ make them understand very quickly that you're not a guy to play around with and to bullshit. And on top of that, yeah, you use technology to drive growth and ⁓ really wake up every morning, see the tangible results of your work, what you have done yesterday, how you adjusted Google Ads or meta ads or a new feature that you have shipped ⁓ on. and now recently we've launched the shopping app, the mobile app, so maybe a new feature that you have shipped on the mobile app and how the customers are actually responding. Are they loving it? Are they hating it? And you know adjust accordingly. Connor (02:35) Is that something you do pretty often kind of get feedback from your users and use them as your, your, maybe your guiding metric. Dmitriy Anderson (02:41) Absolutely. This, as any e-commerce guy would know, this is absolutely critical. have... different ways of getting feedback from the customers. We use Hotjar that is ⁓ integrated directly into our Slack. So as soon as the customer is frustrated, we have ⁓ Hotjar kind of little surveys placed in different strategic places in the customer journey at the checkout as well to ask the customer what happened. ⁓ Are you having issues with signing in to the website, into your account? Are you having issues with checking out and And then we get a message that pops up in Slack and we look at it immediately. We do our best to really reach out to every single customer and try and convert it into a sale. At the end of the day, every frustrated customer is a potential sale, is a potential five-star review. And something that I've been particularly obsessed with is checking reviews on App Store and Play Store on daily basis. I've set up notifications, I get them on my phone. As soon as we get a one-star review or something below four, we reach out to that customer immediately and try to understand what happened. Is it the app that they didn't like? Is it a shitty delivery service that didn't work really well? Anything happens. Anything can go wrong. Connor (04:06) I love that answer. Yeah, we're always trying to push for just more user feedback, more context from the people actually using your products. And it'd be surprised, I feel like lot of people actually don't do that. ⁓ Do you feel like that's maybe a core component of your approach or like, do you think that makes your approach different in any way than other people? Or do you think you're using it in the same way that other people might be using it? User research. Dmitriy Anderson (04:36) Look, I think it's ⁓ a combination of ⁓ factors. We try to bring exciting new features to the South African market. what we try to do is go and check what Amazon is doing, what Best Buy is doing, what Target is doing. Walmart, we look at our competitors in the UK, like BNQ and stuff like that. And then we don't just go ahead and copy paste. We try and create our own kind of mismatch of features. the best and then test it on South African market. And South African market is so young that almost any feature from the first world economies that you bring ⁓ to South Africa, it just explodes. People absolutely love it. They are so hungry and ⁓ it works. And then ⁓ the customer feedback, we just push those features into the feedback loops. the more Connor (05:22) Interesting. Dmitriy Anderson (05:32) the more feedback you get from the customers, the more you can perfect the feature and make it actually useful for the South African people. And that's our main goal ⁓ of what we're trying to achieve to be useful and to, as our tagline says, help them make their home the best place to live. Connor (05:52) Yeah. Do you have an example of that with anything specific or off the top of your head or? Dmitriy Anderson (05:58) The most recent one, we are integrating Uber, for example, and we are going to offer same day delivery or on what we South Africa, we call it on demand delivery. So if you are working on a project and you need, I don't know, some screws or you forgot to buy an additional bucket of paint, you can always place an order and then Uber will just. come and bring it to you in no time. So that's one of the really important features. In South Africa, general, South African e-commerce has been really obsessed with express delivery, one-hour delivery. It's really strong in food retail, but now DIY is catching up. will be the first on South African market ⁓ as a DIY retailer who will implement this thing. Connor (06:47) fascinating. ⁓ Justin Aronstein (06:49) Do you, so you guys do a lot of user research, which well done. Sounds like you're integrating your data. Are you guys running a B tests and experiments as well to measure those new features? Dmitriy Anderson (07:02) We are running like hundreds of them, if not thousands. I don't remember when was the last time I looked into dynamic yield. We use dynamic yield as a... ⁓ know, A-B testing and personalization and recommendations engine. It's a MasterCard thing ⁓ since couple of years. And it allows us not to only do A-B testing super easily without any coding experience, ⁓ like implementing front end features. Obviously not really optimized for performance, but at least it allows us to the idea and ⁓ mixed with MasterCard, card swipe. data from different parts of South Africa, parts of even our neighborhoods in the city and stuff. It's really valuable. It really helps us to get the bank for our buck. I think our returns are absolutely crazy. It probably constitutes at least 25 to 30 % of the total. online revenue. wanted to say one P plus repeats our internal jargon, first party retail plus third party retail. Overall, personalization and recommendation and A-B testing brings a lot of money. That's for sure. And obviously customer satisfaction because you can see which ⁓ Justin Aronstein (08:17) That's awesome. Well done. Dmitriy Anderson (08:25) They call them variation, which variation is actually succeeding, which variation is bringing more satisfaction, more money at the end of the day. And you just leave that. And then you integrate it into the code with the web agency in order to optimize the performance. Justin Aronstein (08:40) That's awesome. That's really ⁓ cool. How do you guys prioritize experiments? What's your framework? What's that look like? Dmitriy Anderson (08:49) To be honest, there is no framework. just try things, know, like creative. We go through popcorn sessions and just do things. You can just do stuff. We just do stuff. We just test. see, ⁓ know, check it out. Amazon has shipped a new thing and they ship things like, I don't know, 50 times per hour. Every time you refresh it, you get a new feature. You get something new, like a new drawer window popping up or some... Justin Aronstein (08:53) you Uh-huh. Connor (09:00) Nice. Justin Aronstein (09:03) Mm-hmm. Dmitriy Anderson (09:18) new, we have, we also have retail media, by the way. ⁓ so we are trying also to copy their retail media, ⁓ and of best practices and what, what they, they are testing. ⁓ so yeah, it's just going out there, getting inspired by what we see on, on X, what other people are sharing, ⁓ and what other big retailers are doing and just trying and copying something works. Some, some stuff doesn't work and it's okay. Justin Aronstein (09:19) Uh-huh. love it. Connor (09:46) Has there been anything from the South African market that's kind of been the reversal of what you've experienced where it's, you know, they're seeing things like the on demand or the buy it today feature that's really, you know, exciting for them. But is there anything that you're seeing maybe they're doing that, you American markets aren't doing? Dmitriy Anderson (10:08) ⁓ I think it's exactly the on-demand thing, which actually has been introduced by a big food retailer called ShopRide Checkers. ⁓ It was during COVID, they have introduced ⁓ this service called Checkers 6060. So basically you can get ⁓ food or some groceries delivered within 60 minutes. And this is something that I'm a New Yorker myself and This is not something that you can always get from every retailer. It depends on where you stay. If you have CVS or Target in your neighborhood, maybe that store supports it, maybe it doesn't. But South African market has implemented this in like 2020. So it's been good six years since this. services available and they actually expanded it. have those little kind of mini, mini electric buses that they use for larger products, not only ⁓ food, but now beyond food and groceries. ⁓ They have gone into FMCG and some even some DIY stuff. So that's, I think that's been the only really explosive thing that they have done and they're quite obsessed with it. So Justin Aronstein (11:06) Thank you. Dmitriy Anderson (11:22) We absolutely have to catch up, but luckily our competitors, Builders Warehouse and others, Builders Warehouse, they ⁓ actually part of Walmart. But luckily enough, they haven't caught up with this idea and we will be first. We will beat them to it and implement it first. Connor (11:42) Nice. Justin Aronstein (11:42) love it. I love it. Yeah. we've actually, we've seen some. Dmitriy Anderson (11:44) You can actually see them in the window. There is a yellow store there somewhere. That's Builders Warehouse. So we can see one of their stores. Just to remind us not to relax. yeah, they're part of Walmart. Walmart actually opened up a couple of weeks ago in South Africa, their first few stores. So yeah, the market is really lively. Amazon arrived in 2024. Justin Aronstein (11:49) ⁓ huh. Okay. Connor (12:04) interesting. Dmitriy Anderson (12:12) ⁓ So we are starting also to catch up with the US in terms of all the different brands available and products and so on and so forth. But we're still pretty far. Still lots to do. Justin Aronstein (12:27) I love it. So it sounds like you've had a lot of success, a lot of growth, learning from other industries, learning from other markets. What's one of your more memorable failures that you look back on as a learning opportunity? Dmitriy Anderson (12:43) Yeah, I think as I shared with you ⁓ just before the recording, the biggest failure I think that ⁓ I actually didn't manage to learn from ⁓ multiple years in a row, unfortunately, like learn from my mistakes is ⁓ overspending on advertising budgets. I think this is ⁓ the biggest thing that I remember, the biggest negative aspect that you absolutely have to learn from, you don't want to learn from it because you do want to sales, you want to drive traffic, you want to drive conversions and eventually success for the business. But ⁓ as we discussed, the weather has been really, really bad ⁓ lately due to natural phenomena. And sometimes no matter what you do with your advertising and no matter how much you overspend, the market just doesn't want to budge. You there is no, there is no elasticity. ⁓ and at the end, you, you sometimes end up overspending without the necessary return or expected return. And then you have a beef with your CFO, with your CEO, with your president and so on. yeah, that's, that's, that's not cool. Justin Aronstein (13:54) And the overspending, was it ⁓ because you weren't getting the top line revenue or was it really a profitability problem and you weren't able to understand the profitability at the time of the spend? Dmitriy Anderson (14:11) Yeah, look, I think the overspend with the latest average real-life figure, updated one from Google Ads, for example, was around 7. So even though 60 million or whichever, sorry, 60 million was the result that I managed to drive and I overspend by 7, 6 or 7 or something, ⁓ it was still not enough from the top line perspective to drive enough bottom line, right? So it kind of the overspend at the bottom outweighs what the top line could bring, unfortunately. And at the end of the day, the result was already negative and I made it worse. So I got into a little bit of a trouble. And yeah, I won't make this mistake anymore. Justin Aronstein (14:41) No. Yeah. And so you had a seven row as though. I mean, most advertisers, you tell them it's a seven rise, they'd be ecstatic. Dmitriy Anderson (15:05) Yeah, that's It's not bad. In 2024, we managed to achieve over 10. I think it was 10.8 or 11. I mean, we know our thing. We know how to do stuff. But sometimes, as I said, seasonality and weather just gets in the way and there is nothing you can do. When you lose hundreds of millions because garden... you know, furniture, patio furniture and all the barbecuing stuff in South Africa, we call it braai. All your braai stuff is not selling because it's raining. People are not outdoors. So there is nothing you can do. So, but you still, you're still trying, you're trying to push other categories. You're trying to push it in Cape Town because there the weather is better, but the market sometimes just doesn't want a budget. It's stubborn. And 2025 was the year of a really, really stubborn market, unfortunately. Justin Aronstein (15:59) Mm-hmm. Love it. ⁓ Yeah, that's a great lesson to learn. Through that, do you have any e-commerce hot takes? Dmitriy Anderson (16:11) Absolutely. Yeah, and I think this is how we got to this podcast when I shared my hot take with you and... It was around AI, ⁓ as I shared with you before, I'm quite obsessed with AI, wipe coding, and how ⁓ even the results of wipe coding can be integrated into the ⁓ actual e-commerce platform and how AI can actually help us drive sales. ⁓ In our case, a couple of years ago, we have partnered with a big famous French AI shopping co-pilot company called iAdvice. They have many big customers in France and outside of France, even in America. And ⁓ we've been kind of fighting with them and driving them to be a little bit more... courageous and bold and more open-minded. I'm using nice words because it's a podcast. And you know, just like we do, stop trying to frame yourself, stop trying to limit yourself because you are trying to be careful and safe. Justin Aronstein (17:15) You Dmitriy Anderson (17:25) ⁓ and just try to do things and see what happens. And we succeeded to a certain extent and we made almost $3 million of revenue last year thanks to AI, driven specifically by the shopping copilot and mostly by the actual ⁓ pre-populated product questions on the PDPs. Exactly what Rufus is doing today. If you check my PDP and an Amazon PDP, they... You could almost say that they are using iAdvice. They have developed their own thing, which is Rufus, which has become super efficient. And I've been reading about the latest data. They actually managed to add additional 60 % or grow their conversion rate by 60 % thanks to Rufus in the past few months, which is absolutely incredible and which proves my point and proves my hot take that customers... ⁓ love when there is something interactive on the PDP. ⁓ When the questions are pre-populated for them and they don't have to think, I mean, everybody can put a, you know, kind of a charge of PT context window on the PDP and say, ask your question. 99 % of customers will ask nothing unless you suggest something, right? You need, you need to reach out and you need to suggest something. You need to suggest some useful questions and then they will start interacting with it. And, ⁓ Justin Aronstein (18:42) Yeah, that's right. Dmitriy Anderson (18:51) In my case, the conversion rate is 2.5 % from all the chats that all the conversations that customers have started, 2.5 % of them have converted, which is absolutely astounding. And ⁓ I believe that the future of e-commerce actually and branded e-commerce websites ⁓ might lie in the entire page being converted into one big shopping Justin Aronstein (19:00) Thanks. Dmitriy Anderson (19:21) assistant or big shopping copilot, like a fully immersive experience, right? Without the traditional search bar, the old useless ⁓ department or catalog menus and all of that boring stuff. ⁓ Rather than that, you are ⁓ greeted by a copilot that is, yes, it is a context window. Maybe ⁓ in the future, we are actually working on that with 11 Labs. ⁓ You can actually chat with it. You should be able to chat with it, like verbally have a conversation, speak to it, and check out by speaking without even typing anything, because some people are lazy and they don't want to type. ⁓ In AI communities, we joke that some people are even lazy to speak to their machine, to their computer. But we need to give them all different... Justin Aronstein (20:13) Mm-hmm. Connor (20:14) you Dmitriy Anderson (20:17) possibilities, all different ways of interacting with the platform. And I really believe that the future can be after such a solution. And I advise that they have actually developed ⁓ kind of an alpha version of that idea where your copilot window expands and takes over the entire website. And only on the right side, on the right pane, you have a little bit of space for traditional banners, for a little bit of retail media. But most like 80 % of your of your page is the copilot and you are literally having a chat. You are not looking at the products you are asking the copilot what you're interested in. For example, one of the use cases was I needed to set for my bike. I would like to fix my bike and I would like a tool set that is specifically ⁓ good or useful for a bike. Try and try and find such a thing through a search bar. Sometimes it can be very difficult. because it's very specific and AI really, really handles it really well and then it offers you additional things and stuff. Justin Aronstein (21:26) I love it. was ⁓ first up, you are the second person to come on the show and talk about I advise it specifically. So they're obviously gaining momentum. I think they have a really good solution. But what percentage of your Connor (21:36) Hmm. Dmitriy Anderson (21:40) ⁓ To be honest, they were ahead of the game and I was a little bit frustrated with them recently telling them that look what Amazon has done. They have released Rufus, ⁓ I think like a year or year and half ago and in the very beginning it was utterly useless. It was just a joke. And now a months after it has become such an incredibly powerful tool. And you guys had like a... ⁓ head start of at least two, three years and you allow them to catch up. And the answer was not really satisfactory to me because they said, you know, they have such big R &D budgets. mean, come on, guys. It's your bread and butter. You do just one thing. So they have to catch up. Justin Aronstein (22:26) Yeah. What percentage of people on a PDP are interacting with it? You said two and a half percent conversion rate. What percentage of people on the product details page are actually interacting with that iAdvise widget? Dmitriy Anderson (22:40) Let's say in a month if we have 4 million sessions, least 250,000, 300,000 of those sessions interact with the questions and with FAQs. And those FAQs get actually the highest NPS score. They always sit above 70. Consistently. Whereas the actual chat box where you can, regardless of where you are on the website, you can click on that little bubble and you can start typing your own question and you can search for something. They call it shopping. I don't remember, founder or product or something like that. Justin Aronstein (23:19) Mm-hmm. Dmitriy Anderson (23:32) And that actually doesn't work and customers don't really enjoy it that much. The NPS is, or CSAT is traditionally low. It's like below 40. ⁓ Surprisingly or not, they don't, again, if when they don't get pinged by the tool and they don't get a predefined question or something to ask or like a conversation starter, it just doesn't work. Because they don't know what to type. They don't want to type anything. It's like, you know, leave me be. I just want to do my shopping, check out quickly. Unless you reach out and you start the conversation. And this is exactly what I've been telling. Justin Aronstein (24:07) Totally. Totally. Dmitriy Anderson (24:15) I advise that they need to ⁓ upgrade the experience and they kind of have done it in our website where when you just enter the website at the bottom, there is like a little pane that pops up and it has a few predefined questions that appear and you can ask your own questions. So it's a little bit, it's one step further from just having the bubble, you know, the typical chatbot bubble. Justin Aronstein (24:40) Yeah. ⁓ I love it. actually think why don't I think that's a test that you could build though, using the, advise APIs to then build it in dynamic yield to call those APIs and JavaScript build that thing. And not using what I advise has. Dmitriy Anderson (25:00) Exactly. You can do crazy things and maybe even reach out to Amazon and say, we, do we have Rufus APIs? I mean, come on. Justin Aronstein (25:06) Yeah. I love it. love it. ⁓ Awesome. So what's the structure of your team? ⁓ How big is your team? What kinds of things are they working on? Dmitriy Anderson (25:25) The team is very small. have around five people working on first party e-commerce and I have a team of eight boys and girls working on the marketplace, on the business side of the marketplace, onboarding sellers, managing the range of products of about 150,000 unique SKUs. And in our company and even in a broader group overall over the world, famous for that, having the ability of running big businesses that make a lot of money with small teams. And I'm a big, big fan of that. ⁓ I don't believe in overblotted teams unnecessarily. Obviously, we work with some agencies. have a web agency, Magento agency, that is helping us. ⁓ build and maintain the website. I have a mobile app agency that works out of India that helps me with the app. So I outsource a lot of tasks, but the internal team is very small. And that brings us a lot of productivity and lot of fame inside the group. Because e-commerce teams are traditionally really huge. You have a couple hundred people doing only God knows what. Justin Aronstein (26:46) Sure. How do you know who to keep in-house versus when to hire an agency? Connor (26:47) Mm-hmm. Dmitriy Anderson (26:55) ⁓ traditionally I strongly believed in not keeping developers in house because it's really difficult to, to keep them overall, ⁓ in a, in a DIY, ⁓ hardware, ⁓ retail is just something really, really boring. There is not a lot of growth for them internally because it's, it's at end of the day, it's not a software company. It's a retail company that happens to make software because it's a need. ⁓ right. And, ⁓ in the early stages of my. career or experience at Leroy & Merlin. I tried ⁓ hiring developers, tried hiring data guys, BI guys, but as soon as they grow, as soon as they gain experience and reach a certain level, they become bankable and they just get poached away very, very swiftly. So it's kind of a zero sum game because you invest into those people, you train them. ⁓ Connor (27:43) Mm. Dmitriy Anderson (27:50) you send them to for external trainings, you invest into their certifications, and then they just leave. Justin Aronstein (27:59) I think everyone has that problem. And think retail and e-commerce, there's just such an array of skills that are necessary. You need customer service, you need sales, you need someone in the warehouse, you need the marketers, you need the product managers in terms of like, how do the products are, all the information and sourcing the products. And then you need the technology arm and Connor (28:01) Yeah. Dmitriy Anderson (28:12) Yep. Justin Aronstein (28:29) Creating a culture that these developers want to work in while also catering to a culture of this larger company is so difficult. I'm not going to say impossible, but it's really, really difficult to do. Dmitriy Anderson (28:33) Exactly. Exactly. And that's why bigger, larger enterprises or companies like Walmart or even like ShopRite Checkers in South Africa, Walmart has created Walmart Labs. In ⁓ ShopRite Checkers, have created... ⁓ Justin Aronstein (28:54) Yep, that's right. Dmitriy Anderson (28:59) X Labs or something like that where it's a dedicated standalone business unit that is investing in the technology, dedicated to technology, then it starts making sense, right? Where you can excite the developers, the technical people, the data people with a lot of fun because you have funds to have fun. ⁓ When it's small company and unfortunately you cannot afford that and it doesn't make sense from a business development perspective, you would rather invest into opening new hypermarkets in different cities expanding nationwide rather than building these kind of labs just to have fun. So yeah, we have to focus on business and making money and here we are doing exactly that. Justin Aronstein (29:47) That's right. Is there anything you wish the executive team knew about your job that they don't quite understand? Dmitriy Anderson (29:59) No, to be honest, I'm really happy with them not understanding anything about my job. ⁓ I think it's the best place to be. And ⁓ as they say in AI communities that ⁓ the less defined your job is, the more secure it is in the future. Yeah. So the job spec has to be ⁓ vague, undefined. Connor (30:05) You Interesting. Dmitriy Anderson (30:25) Nobody knows what you're doing. You're like a magician, you know, using your dark magic, making things happen, making revenue appear and everything grow without really understanding what you're doing. So let's keep it this way, please. I'm happy with it. Justin Aronstein (30:43) it. What are you reading, listening to, watching to kind of prepare yourself for 2026 and any trends that you're watching for this year? Dmitriy Anderson (30:56) ⁓ Yeah, lately I've been watching a lot of stuff on Cloud Code and wipe coding and ⁓ especially in the past few, maybe a week or couple of days, I've been focusing on this new methodology of coding thanks to Cloud Code called Ralph Wiggum, which is basically ⁓ bash loops where you give it a task, you know, before going to sleep, you wake up in the morning and your application is fully developed. Connor (31:16) from the city. Dmitriy Anderson (31:25) And it keeps going in loops and runs many, many iterations until it perfectly completes the task that you gave it in your MD files. As we say, who knew that we were going to be coding in that MD? So I'm watching only that stuff, like obsessed. Can't sleep. Justin Aronstein (31:48) Wow, I have not watched any of that. Now I'm really interested. I'm gonna have to go pull some of that up. Dmitriy Anderson (31:52) You have to, especially with your engineering, software engineering background, you must. It's a sin not to try. Justin Aronstein (32:01) Yeah. Awesome. Well, thank you so much for your time and your insights. I think this was absolutely perfect. And thank you so much. Before you sign off. ⁓ but. Dmitriy Anderson (32:13) Thanks so much. No, I was saying it was my honor. It was awesome to participate and have a nice chat. Tell you a little bit about South African market. Yeah. Connor (32:27) Okay.