Listen
Watch
See Boozt’s fulfillment center with over 1,000 robots that augment their team to deliver orders quickly and efficiently.
It’s not every day I get to see 1,000 robots all in one place, so today is special in a few ways. First of all, I’m here in Ängelholm, Sweden at the Boozt Fulfillment Center, and am about to take a tour of their facility – robots and all – in a few minutes.
But before that, we’re going to talk about creating amazing efficiencies while also delivering on an amazing customer experience in online retail. After all, why can’t you have a lower cost per transaction while raising your net promoter score?
To help me discuss this topic, I’d like to welcome Tobias Sjölin, CTO of Boozt Fulfilment & Logistics – a division of Boozt Group.
Resources
Boozt website: https://www.boozt.com
Wix Studio is the ultimate web platform for creative, fast-paced teams at agencies and enterprises—with smart design tools, flexible dev capabilities, full-stack business solutions, multi-site management, advanced AI and fully managed infrastructure. https://www.wix.com/studio
Attend the Mid-Atlantic MarCom Summit, the region’s largest marketing communications conference. Register with the code “Agile” and get 15% off.
Register now for HumanX 2025. This AI-focused event which brings some of the most forward-thinking minds in technology together. Register now with the code “HX25p_tab” for $250 off the regular price.
Connect with Greg on LinkedIn: https://www.linkedin.com/in/gregkihlstrom
Don’t miss a thing: get the latest episodes, sign up for our newsletter and more: https://www.theagilebrand.show
Check out The Agile Brand Guide website with articles, insights, and Martechipedia, the wiki for marketing technology: https://www.agilebrandguide.com
The Agile Brand podcast is brought to you by TEKsystems. Learn more here: https://www.teksystems.com/versionnextnow
The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op. From ideation to creation, they craft human connections through intelligent, engaging and informative content. https://www.missinglink.company
Transcript
Greg Kihlstrom:
So it’s not every day that I get to see over a thousand robots all in one place. So today is special in a few ways. So first of all, I’m here in Angelholm, Sweden at the Boozt Fulfillment Center and just took a tour with Tobias of their facility and got some pictures, some video. We’ll share that on the show as well. But now we’re going to talk about creating amazing efficiencies while also delivering on an amazing customer experience in online retail. After all, why can’t you have a lower cost per transaction while raising your net promoter score and your customer satisfaction? To help me discuss this topic, I’d like to welcome Tobias Sjölin, CTO of Boost Fulfillment and Logistics, a division of Boost Group. Tobias, welcome to the show. Thank you very much. It’s a pleasure to be here. Yeah, I’m really excited to be able to be here in person. So why don’t we get started with you giving a little background on yourself and your role at Boost.
Tobias Sjölin: Great. So I’ve been at Boozt for 13 and a half years now. So you should say it’s me and the plastic flower still here. So we have a group of people that’s been there for a very long time. But I’ve been part of the company when it was very, very small. We had very manual warehouses and processes around it. And now we’re sitting in these fulfillment centers with 86,000 square meters and robots all over the place. So my job in all things, as I said, I’m CTO for this specific Boost Fulfillment Logistics company. So my responsibility is to make sure that all the software systems are working as efficiently as possible, but also in collaboration with all the hardware that we have as well. So as you mentioned, we have hardware from a company called Autostore, which is Norwegian. So then we try to make this as efficient as possible with software, basically. So you have three different teams with developers working with software. Great.
Greg Kihlstrom: And for those that are not as familiar with Boozt, can you talk a little bit about what is Boost and what do you do?
Tobias Sjölin: Sure. So Boozt is an e-commerce retailer in the Scandinavia. We have a Scandinavian focus. So what we do is we focus on something called Nordic Department Store. So we don’t only have fashion, we also have interior design, we have sports, beauty and so on. So it’s basically a department store for the whole family. So what we do is, as I said, we focus on Scandinavia. Our biggest markets are Sweden and Denmark. We have a European presence as well, but we try to focus on the markets that are close to heart. So that’s why also the Fulfillment Center is here in the southern part of Sweden, because then we can reach all of our markets very, very efficiently. So basically that’s what we do. We get items into the warehouse and we send them out again. And then we have a whole headquarters in Malmö as well. It’s on our cell phone here, where we have our own marketing team, our commercial team, design team, customer service, financial team, and so on. So what you see here now is the fulfillment process of the company. But then we have another 500 people in Malmö and then some other development centers as well.
Greg Kihlstrom: Great. I want to start by talking about the online retail customer experience. And so, as I just saw, taking a tour, you know, very efficient and just massive. Again, we’ll have the video in the show notes as well. But, you know, just a lot of moving pieces here and a lot of customers that are expecting, you know, they have high expectations from their orders. So you’re mostly online, right? So I know you have a couple retail stores, but primarily online. So it’s incredibly important that the e-commerce experience is great. How does Boozt approach that from a staffing and a team perspective?
Tobias Sjölin: I mean, everything we do is from a customer point of view. So when we walked around here as well, I was mentioned that when we do different automations or different process changes, we always think about the customer first. So is this improving the customer experience? Is this making delivery faster for a customer or in any way better for the customer, right? So it goes through all of the different teams. So when we do from a commercial point of view or different campaigns and stuff like that, we always focus on the customer first. So what we do when we do different prioritations, for instance, so one thing is from business value, but mainly we discuss what is the best for the entire company, and that is usually what is best for the customer. So what is best for the end customer? And that is what has the highest priority. And then we work our way through into the whole. So we have different roadmaps all over the place, but in the general strategy, that’s how we work with. So our customers know that’s how we make our money. I think you mentioned we have two different physical stores as well. I think those are less than 0.01% of the revenue. So what we live by is our revenue from our online store. So if the customers are not happy, then we don’t get any money.
Greg Kihlstrom: Yeah, and so with that focus on the e-commerce experience, how much of what you create is customized from a technology standpoint, and how much is off the shelf, so to speak, from a technology perspective?
Tobias Sjölin: So almost nothing is off the shelf. We have 210 developers working in different teams, so around four to six developers in each team. And each team has a responsibility. So you have one team responsible for checkout, one team is responsible for the robots, one team is responsible for listing page, and so on. So you have all those systems are custom made. So we started with a web shop, and then we’ve done our own content management system. We have our own CRM system. We have our own warehouse management system, warehouse control system, ERP system, and so on. So the whole core universe of our e-commerce is made by ourselves. That’s also what makes it fast. We can do very, very fast changes. And we talk about processes, how to improve them rather than different bottlenecks in a system. So we never say, oh, the system can’t do that because of X. It’s always like, oh, what is the most efficient way of doing this? And then we change.
Greg Kihlstrom: That also seems important because a lot of e-commerce experience feels kind of cookie cutter. Like you go from one store to the other, it all feels the same. So it seems like that’s… a differentiator as well, right? We hope so.
Tobias Sjölin: I mean, especially we try to, one of our main advantages is our fast deliveries. So we try to keep that next day delivery. Even when we were, so 13 years ago when I started, so even then we focused on fast deliveries. And then the problem has been when you scale a very big e-commerce, how do you keep the speed of the whole thing? Because as I said, we have 800,000 different items in the warehouse, right? So now we need to find one of them. way longer time than what was before, right? So that’s a constant kind of evolution of, you know, how can we just make it faster and better and make sure that we reach the customer in time.
Greg Kihlstrom: So what kind of measurements and methods do you use to make sure that you’re optimizing and having that optimal customer experience?
Tobias Sjölin: I don’t know if all of them is a good answer, but that’s kind of how it feels like, right? So, I mean, of course, I mean, we use Google Analytics to make sure that, you know, the website is working as it should. We do different A and B testing on that one. But then we have monitoring on all of our robots, for instance. We have, you know, we measure battery quality on every single robot, you know, and then we measure, you know, everyone who is out there picking and packing, we measure, you know, is something item going wrong, going out, and so on and so on. We tried to be on a very, very low level of monitoring and gathering data. So one problem we had, we actually ran out of Google Analytics transactions. So we sent too many transactions to Google Analytics. Yes. So we had to find other technologies in order to keep up with the data amount that we’re sending them. So we do that a lot.
Greg Kihlstrom: What role does self-service play in the end customer experience? And I believe that self-service also applies to internal teams as well.
Tobias Sjölin: Yeah, so I mean, we have something we call MyBoozt, for instance. So you log into a site and when you can see your orders, you can see, okay, where’s my refund, you know, or return or orders and so on. And then you can also see unpaid invoices and, you know, different things like that. So it’s important for us that customers should have their own tools to do stuff. So we have a saying in the Plasma team where I’m at, it’s power to the people. And the people changes sometimes. So people can be end customers. But it can also be internal business units, so like the commercial team. They shouldn’t have to go to us to ask to change something if they can do it themselves. So that’s a whole self-service thing that we try to do. And also from an internal development point of view, so our developers, they shouldn’t have to go to an engineering team to ask for database access or to deploy something, things like that, to do changes. So they should have all the tools themselves. And then we’ll just trust the people to do the right stuff with that thing. Yeah, that’s great. It’s a cultural thing more than anything else, but it’s important for us to give that responsibility.
Greg Kihlstrom: Yeah, I mean, it seems like from a cultural perspective, it seems like if people are more empowered to be able to do more, they’re going to have more value in their jobs and give a little bit more.
Tobias Sjölin: Yeah. We actually expect a new developer, we expect them to take down the site within the first week. So that’s kind of, you know, it is expected that things will go wrong. Yeah. So what we’ve done is, you know, enforce the whole thing. If something goes wrong, we can fix it very, very fast instead. So we’re not trying to avoid mistakes. It’s rather when something happens, we can fix them very, very fast instead. So we try to minimize the risk of mistakes rather than avoiding them completely. Yeah. Which is very important for us.
Greg Kihlstrom: So now I want to get to the robots. So as I mentioned, I got a tour here, got to take a look for myself. You mentioned AutoStore. So I had a guest from AutoStore on the show a few months back, which is actually why I’m here and how this whole thing came to be. So thank you, AutoStore. So he mentioned that they had this facility in Sweden with over 1,000 robots and just had to see it for myself. So here I am. So I believe you started using AutoStore around 2016 or so as well. And to start, can you briefly describe what exactly is AutoStore for those that didn’t catch the previous episode?
Tobias Sjölin: So basically, AutoStore is like a matrix of different plastic bins. So it’s like a cube of plastic bins standing on top of each other. And then you have robots riding on top of them. and their robots are picking up bins and delivering them to an operator. And then you have an operator and they’re taking the items out of the bins and making sure they put in the right bag and so on. So it’s a dense stock capacity so you can fit a lot of items in a very small space and you still have a speed of delivery when it comes to items to an operator.
Greg Kihlstrom: So, I mean, some of this I think you’ve covered a little bit already, but how do you look at a system like AutoStore being able to help you achieve that greater efficiency and customer experience?
Tobias Sjölin: So, I mean, when we look at a system, when we get hardware to play with, it’s maybe a weird way of putting it, but that’s how we see it. It should be something that we can incorporate in all of our processes, right? So, also help in that sense that we can dive fairly deeply into how they’re working in terms of technology. So from our point of view, the way we use it, we try to keep up the speed as much as possible for the customers again, right? And then we try to see how we can tweak different things. So we have one solution, you get the out of the box, but then we try to see how can we incorporate more things around it? How can we build our processes, either change our processes around AutoStore, or how can we make AutoStore change according to what we want it to be? So a very, very tight connection with AutoStore. with the collaboration to try to find out the best e-commerce processes going forward. So we help them to spread different things how we do it, and then they help us with different technology. So it’s a very nice collaboration for us to do.
Greg Kihlstrom: And so you mentioned in another interview that you did about cost per transaction, and that’s definitely a key way to look at efficiency, and it’s a way that you look at efficiency. What are the factors that go into this?
Tobias Sjölin: I mean, you have, I mean, that’s how we measured all of us from a fulfillment point of view, right? So how much money do we pay per order going out? And then also you can add and we have returns coming back in and then inbound. So everything has a cost. Each transaction has a cost. So you try to either minimize that or optimize it. So you have a robot doing it instead. But even if you have a robot doing it, you still want to make it as efficient as possible because it’s still a resource game. So you can’t just be complacent with it. You still need to be very, very accurate on how you’re doing and how to make it more efficient. So it’s a lot around, I mean, cost for us is the number of touches of an item. So we would like to place an item in the perfect place in the first go. If we have to move it seven times, that’s a cost for us. So we measure every time we move something. We measure how do we put it in a bag and things like that, and how do we make it very, very efficient for all the people. So we have some amazing warehouse people working in the warehouse. And what we try to do is to make them as efficient as possible as well. So all the technology we have here is to make the human we have here more efficient and to do more efficient work.
Greg Kihlstrom: That’s great. And that’s a great way of looking at technology. With all the talk about AI and everything these days, I like to look at it as augmentation, not replacing humans.
Tobias Sjölin: Not at all. We have the same thing. I mean, AI is a great tool. It’s like a tactical tool for doing different things. But it’s not a replacement of everything that it sits in right now. So you have AI, AI can do fantastic things. We would like to see that that’s one of the tools that we’re using to make things more efficient, right? And then you have other things like database optimizations that, you know, no one asks you, oh, how many databases do you have? Because that’s like a void question now. How is everyone doing databases? So for me, that would be the same thing with AI like 10 years from now. AI is something that exists everywhere. This is a very, very powerful tool to use. But it’s not everything. It’s, yeah. Something to elevate humans.
Greg Kihlstrom: Yeah, absolutely. And so, you know, it’s September right now when we’re recording this. We are, you know, getting towards peak holiday shopping season. So, you know, we’re talking a lot about efficiency, a lot about cost per transaction. You know, what does this system look like at its peak?
Tobias Sjölin: So I don’t want to stress out any logistic colleagues I have now. We started planning in May, or peach. So we’re in the middle of our peak planning now. So now it’s like game time. Now it starts to ramping up to the whole thing, which is, so I think it’s the 29th of November, which is Black Friday this year. So everything we do here now is we, So usually what we do is Q1 and Q2, we do new products. And in Q3 now, then we clean up all those products. Then we focus on performance. We focus on the bug fixing, all of that. So all of that, we’re pressure testing the whole warehouse now to make sure we find all those things that will explode during peak. So we’d rather find them now than find them during peak.
Greg Kihlstrom: OK. So you’re planning the whole year basically for Black Friday. Yeah. That’s our Super Bowl.
Tobias Sjölin: That’s, you know, in one go, everything is happening at one time. We also have the same thing with the web chart, for instance. We’re actively trying to take that down with traffic now. So we’re putting in as much traffic as we can until the web chart breaks, and then we roll back to traffic again and say, what break? Ah, there was this one here, and then we can fix that one, and then we push more traffic. So we try to push it. We’d rather break it down now than during peak. That’s the main plan and the whole thing. But during peak, if you’re asking transactions, so we do 220,000 pieces in 24 hours. From this, it’s like 70,000 orders, something like that, in one day, so 24 hours. And that’s kind of… So problem from a fulfillment point of view is that peak is… lingering on for another two weeks. So you have Black Friday where everyone is focused on, especially from a marketing and commercial point of view. But then you have fulfillment for two weeks. We need to get all the orders out. And after those two weeks, then you get the returns coming back in. So it’s kind of a long process for the whole thing.
Greg Kihlstrom: Yeah, it’s not like a day or a few days or something. It’s like a month.
Tobias Sjölin: So we usually see it more as a marathon rather than a sprint. It’s like, OK, we need to just be consistent, sustainable in our operations.
Greg Kihlstrom: So last thing I want to talk about is a little bit more about the digital experience. And we touched a little bit on AI and the idea of augmentation and things. But what are some of the key ways that you’re using AI?
Tobias Sjölin: From a fulfillment point of view or from a general?
Greg Kihlstrom: From a customer perspective, let’s say.
Tobias Sjölin: Right. So we have a data science team in Aarhus in Denmark. So they’re doing everything with machine learning and AI and stuff like that. So they’ve given us some great tools around it. So one is that we have size recommendations, for instance, that we do. So when you’re a new customer, especially going to the site, we can recommend you a good size of the different items. And then we’ve just launched product descriptions now, so for certain categories. So more to elevate the description more than create it from scratch, but more like put some more flavor into the whole thing and maybe digest it from different other internal sources that we may have. So take a picture, for instance. to look at the picture and put some details on the whole thing into the process and so on. And then, I mean, from a fulfillment point of view, we also have different around resource planning and things like that that we try to do. So we try to, it’s very complicated to have 470 people that you’re trying to organize in a perfect way, right? So you have too many people, then the cost increase. If you have too few people, then the delivery times increase. So you need to balance that every day, you need to balance it in a perfect way. But then we also have some around image recognition, again, so for categorizations on the site. So instead of having people sitting in categories, okay, so this is a shoe, and this is a t-shirt, and so on, then we can, you know, just take the picture, and then we can categorize it based on AI, and then we can just do that, and you have people approving it instead, which is way faster. So you just click it, so it gives you a suggestion, and you
Greg Kihlstrom: So there’s like a human review process, but it massively speeds that up.
Tobias Sjölin: Exactly, because you can maybe do 100 instead of one.
Greg Kihlstrom: Yeah, yeah. That’s great. What about internal use of AI?
Tobias Sjölin: Well, I mean, our developers, they love Copilot, for instance. So when you’re coding, Then you get this huge help from Copilot. So basically, when you’re writing a code, you get the code segment suggested for you. And then you, of course, you look at it and see if you want to use it or not. But it helps in different, you know, the simpler tasks. When you’re just writing, you know, the basic code of the things, it can go that very, very fast instead. So usually when we say that we don’t do AI, in that sense, AI is ingrained in everything we’re doing. And it comes mainly from the grassroots. I want to try and co-pilot now. And that’s spread across the whole development team. And then there are other things like from a content management point of view to when we look at the web shop, right? So then we have 800,000 different items, right? And then you want to run through different discount settings and things like that. So it’s hard to go through all those 800,000 items. So then we’d run different scripts and stuff on it to see, OK, is this item on the right discount here? Or should it be lower or higher? Or can we automatically adjust the whole thing as we go forward?
Greg Kihlstrom: Well, Tobias, thanks so much for joining today. Got one last question before we wrap up here. So you’ve given a lot of great insights and advice already. For those companies out there, I mean, again, massive efficiency and operational things going on here. There’s a lot to learn and share. For companies that are trying to be more efficient, what’s a place that you’d recommend that they look first to make improvements?
Tobias Sjölin: Yeah, I mean, there’s a classic saying also, there’s something called walk the floor. So it’s always very, very good to know your processes from a deep point of view. So we talked about it also when we’re out with the robots, right, too. It’s very good for, I have my developers, they’re out picking and packing and they’re doing inbound returns, I think, just to understand the process. So the more you understand, the better you can identify bottlenecks and so on. But then you have, I mean, From our point of view, I think, from Boozt, we’ve benefited a lot to have our own development team as well. And I’m not saying to hire 500 developers immediately, but start maybe, you know, one, two people, something like that, and then see you can do different tweaks and stuff like that. not replacing your big third-party systems, but start with small different tools maybe, helpers, just get them more efficiency, and then you can expand as you go forward as well. Start small and then scale up. It’s usually not a good idea to go big immediately.
Greg Kihlstrom: I’m definitely a fan of the Agile iterative approach. Well, again, I’d like to thank Tobias Sjölin, CTO of Boozt Fulfillment and Logistics, a division of Boost Group, for joining the show, and for inviting me to the Fulfillment Center here in Engelholm, Sweden. And, tack så mycket, is that right?
Tobias Sjölin: Väl så gud, yeah, perfect Swedish. Thank you so much. Very nice to have you here as well, and to get to show you to robots.